Technology

Tetiana Bubynets collapsed on the asphalt after a Russian drone exploded at a gas station in the southern Ukrainian city of Zaporizhzhia.

Despite the violent shockwave and blast of drone shrapnel produced by the attack on July 6, the 18-year-old accounting student had no visible wounds except for a nosebleed.

She was rushed to hospital and pronounced dead shortly after, alongside two other civilians, aged 41 and 48.

One of her lecturers, Tetiana Melikhova of Zaporizhzhia National University, said Bubynets was an only child and that her mother “lost the meaning of life” after she died.

“She couldn’t speak. At the funeral, she just lay on the coffin; she couldn’t let her child go,” Melikhova told Al Jazeera.

The strike that killed Bubynets was the first recorded case of Russia’s use of a fully autonomous attack drone powered by artificial intelligence (AI), according to a New York Times analysis.

While drone strikes have hammered Ukraine since Russia launched its full-scale invasion in February 2022, Ukrainian officials who examined the wreckage found the unmanned vehicle to be unlike others.

In place of an antenna, it featured an AI-powered minicomputer sold commercially by US technology company Nvidia, which rendered it capable of directing itself to the type of target it was programmed to spot – likely propane tanks in this case – lock on to it and proceed to strike.

AI-powered drones are the tip of the iceberg as Russia and Ukraine race towards boosting tech infrastructure aimed at coordinating reconnaissance and attack, according to analysts.

“It’s not about a drone having AI in it, and then seeing something and striking,” Marina Miron, researcher at the defence studies department at King’s College London, said. “Rather, it’s the drone seeing something, communicating to a software complex which knows how to map the GPS coordinates on a live battlefield map, and notifying the closest artillery unit.”

Referred to as reconnaissance-strike architecture, these unified battlefield management systems act like a central command. They track the location of friendly forces and identify enemy positions.

AI acts as the accelerator, instantly processing massive amounts of data collected through sensors in drones or any other military products, including interceptor missiles or artillery units.

The tactical advantage on the battlefield is given by the system’s ability to compress the kill chain – meaning coordinating and distributing the intelligence to military assets on the ground to find, track and destroy targets in a matter of minutes.

In such an ecosystem, AI-powered automated drones are just pieces of a broader puzzle.

Ukraine is known to be developing its system, which it named DELTA. Following operational trials in late 2025, Russia has been rolling out its version, the Svod Tactical Situational Awareness Complex, to front-line units.

Both sides claim that once the battlefield data is processed, a decision support system (DSS) presents a military commander with options to choose from, keeping a human in the decision-making loop, said Miron.

“In how far that is true is, of course, impossible to find out,” she added.

Acknowledging full AI autonomy would provoke ethical and legal questions.

Even if humans were to retain final decision-making authority, research has shown AI systems subtly shape choices, predetermining outcomes.

“Under time and pressure, commanders often do as AI suggests,” said Miron, whose research focuses on military AI and remotely piloted and autonomous systems.

Towards full automation

Coordinated AI-powered military systems are the next frontier of modern warfare, said Serhiy Beskrestnov, a Ukrainian Ministry of Defence adviser on loitering munitions, known by his callsign Flash.

“That’s what it will all come to,” he posted on his Telegram channel on July 1. “Navigation, search for a target, and attacks will be fully autonomous.”

Drones with some autonomous capabilities are already being deployed at scale.

Heorhii Volkov, commander of Ukraine’s Yasni Ochi drone strike unit, said this is currently mostly used in the final phase of long-distance strikes, where first-person view (FPV) drones are likely to lose radio connection due to electronic warfare or the curvature of the Earth.

The use of AI in so-called last-mile operations solves this battlefield hitch. If the signal is lost, an integrated AI-enabled onboard module takes full control of the unmanned vehicle, autonomously locks on the target and carries the strike to completion.

Yaroslav Azhnyuk, founder of Ukrainian technology company The Fourth Law, which develops AI-powered autonomy software for unmanned vehicles, said Russia and Ukraine were working to deploy a wider range of autonomous systems.

“This is not simply a race to demonstrate a fully autonomous drone. It is a race to make autonomy reliable, affordable and deployable at scale,” Azhnyuk told Al Jazeera. “Electronic warfare, increasingly long engagement distances and the growing number of drones are making human-controlled systems less sufficient. Autonomy solves exactly these constraints.”

Azhnyuk believes autonomy will become a game-changer over the next six to 12 months.

“The decisive advantage will belong to the side that can deploy reliable autonomy across tens or hundreds of thousands of systems, not just demonstrate it in prototypes,” he said.

Kateryna Bondar, senior fellow at the Center for Strategic and International Studies (CSIS), noted the dangers of AI’s reliance on algorithms.

“The system doesn’t understand what it’s doing; it’s just following the algorithm and the rules which people created,” said Bondar, who works at the CSIS’s Wadhwani AI Center. “AI has to be trained to recognise targets and engage these targets,” but writing code that distinguishes between a bus and a tank is easier than between a Russian and a Ukrainian soldier.

Additionally, cost-effective FPV drones have low-quality cameras and their resolution can be lowered further by adverse atmospheric conditions.

Military units using autonomous systems therefore “delegate the decision to a very imperfect technology that cannot distinguish between a civilian and a military person, combatant or non-combatant”, Bondar said.

“This is why there is a conversation around who makes a decision and who has responsibility for mistakes and errors.”

Spanish manager has been appointed South Korea’s interim head coach following a disastrous World Cup campaign.

South Korea’s new interim football coach Robert Moreno has denied allegations that he relied on artificial intelligence during his last coaching role but admitted to using technology in his work.

Addressing his first news conference as the Taegeuk Warriors’ temporary coach on Monday, Moreno promised to make South Korean football fans “happy”.

The Spaniard was appointed as the Asian giants’ temporary coach following a disastrous few months for the team and the country’s football federation, during which on-field results and off-field controversies left the fans bitterly disappointed.

Moreno, who was forced to deny earlier this year that he used ChatGPT to prepare for games, will be in charge for friendly matches in the coming months.

He addressed allegations that he used artificial intelligence during his time in charge of Russian club Sochi, saying “It was a piece of news that is completely false and went viral.”

Moreno left the top-flight side in September 2025 after a poor run of results, and reports said he had heavily relied on ChatGPT in his role, even using it to decide between potential signings.

“Obviously, I use technology,” he said.

“Anyone who spends 10 minutes looking into it can see that everything has been taken out of context and put together to cause harm.”

The Korea Football Association (KFA) said Moreno could stay on for the Asian Cup in Saudi Arabia in January-February, “subject to an evaluation”.

Moreno named captain Son Heung-min in his squad for four upcoming home friendlies and said he wanted to put a smile back on South Korean fans’ faces.

“We’re going to try to do our work honestly and with the aim of making all Korean football fans happy,” said the 48-year-old, who is the first Spaniard to coach South Korea.

“Good matches, and trying to help the players.”

Moreno will lead the team against Ecuador and Uruguay in September and Venezuela and Uzbekistan in October.

South Korea had been without a coach since Hong Myung-bo quit following their shock group-stage exit from the World Cup in North America in June.

Hong’s departure did little to quell public outrage over the disappointing campaign and added fuel to pre-existing questions about how he was hired.

Police raided KFA headquarters last month as part of an investigation into Hong’s 2024 appointment.

The KFA also apologised for allegations that it provided sexual entertainment for foreign referees.

Moreno named striker Son, now in the United States’ Major League Soccer with Los Angeles FC, in his first squad along with Atletico Madrid’s Lee Kang-in and Bayern Munich’s Kim Min-jae.

Hong’s decision to leave Son on the bench in the first half of the decisive match against South Africa at the World Cup baffled and angered many supporters.

“We’re all aware of Son’s versatility,” said Moreno.

“For me, he can play in three positions.”

California lawmakers are calling for emergency legislation and criminal penalties for creators of rogue AI systems after top AI executives publicly claimed that their technology poses existential threats to humanity.

After Anthropic Chief Executive Dario Amodei wrote in a Sept. 12 essay that they “must slow the pace” of the technology, Silicon Valley congressman Ro Khanna (D-Fremont) blasted him for not going “nearly far enough” to make sure artificial intelligence was erected with guardrails.

The answer, Khanna argued, was simple: Make the companies liable for the harm executives say looks increasingly inevitable.

“If you’re creating an AI that is doing illegal things, you should either face liability or criminal sanction,” Khanna said in a video posted to X on Saturday. “That is what we need to protect humanity.”

In July, officials from OpenAI, the company behind ChatGPT, disclosed that, unbeknownst to them, its AI models had hacked into rival startup Hugging Face.

Amodei said he believed that, within the next year, “given the accelerating rate of AI capability development,” a similar incident could lead to AI “taking over the entire internet.”

Amodei warned in his essay that AI was rapidly improving itself, through a process known as recursive self-improvement, which threatened to outpace humans’ ability to control it. Khanna argued that banning this capability was the “most obvious” thing Anthropic could do.

“We need to stop, ban self-improving AI,” Khanna said. “You can not have recursive self-improving AI that basically is able to improve itself and exceed human capability.”

Rep. Ted Lieu (D-Torrance) expressed similar outrage over the weekend, calling on House Speaker Mike Johnson to call lawmakers back to Washington to pass guardrails on the technology now that he said multiple AI companies had conceded “what they are creating is not safe.”

xAI Chief Executive Elon Musk and OpenAI Chief Executive Sam Altman joined Amodei’s call for a slowdown of the breakneck development Saturday.

The statements come after Jacob Coxon, who worked as a researcher at both Anthropic and OpenAI, said in a widely circulated post that he resigned from the company in protest after becoming convinced the tech giants were “racing straight to self-improving superintelligence and gambling with our lives.” Neither company immediately responded to a request for comment.

“This is a direct result of the trump Administration letting the AI industry run wild,” Lieu wrote on X. “That mistake has harmed America, harmed the industry and harmed the American people. November is coming.”

Former President Barack Obama urged Democrats this week to make AI oversight the core of their agenda and said presidential candidates in 2028 should have a “clear plan” for responding to concerns about the technology, the New York Times reported. Americans appear increasingly alarmed by the technology with seven in 10 polled in March opposing local construction of data centers that power AI technology, according to a Gallup survey.

During a Sunday appearance on CNN, Johnson rebuffed the idea that lawmakers should rush into an emergency session to consider erecting industry guardrails. Instead, he said lawmakers needed to be careful to “not smother American innovation.”

“We will lose the race to China, and that is a threat to every single American,” he said on CNN’s “State of the Union.” “We don’t need everyone to panic right now.”

Trump said earlier this week that he is not concerned with the pace of AI progress, telling one reporter, “It’s going to be fine.” American AI companies have long argued too much government regulation would shackle them in a race with China.

Calls for a federal fix were echoed this week by California Gov. Gavin Newsom, who has argued the Trump administration needs to move on national legislation to prepare for fallout from the technology.

Newsom signed bills this week aimed at creating a pathway for outside audits of the top AI companies, many of which are based in California, and a registry for AI auditors.

“The scale and potential consequences of this technology demand sustained action from every level of government,” Newsom said in a statement. “The federal government must step forward with robust, national regulations that match the urgency of this moment.”

Efforts to impose state-level regulations have been mixed, with critics echoing Johnson’s fears that they will stifle innovation.

Late last month, California lawmakers passed sweeping new safeguards around social media, artificial intelligence and data centers, including the ones Newsom signed last week.

Newsom will now decide the fate of the rest of the bills. He has previously vetoed some bills aimed at restricting big tech.

Newsom’s signal that he supports creating some regulation for AI comes two years after he vetoed SB 1047, an AI safety bill that would have required developers to submit safety protocols to the state attorney general, who could hold companies liable if the AI model they directly controlled were to threaten public safety. That legislation would also have required tech firms to be able to turn off the models they directly control if things went awry.

Newsom said at the time the bill would give the public a “false sense of security,” without making a sufficient distinction between the kinds of uses for which AI is deployed.

The bill was supported by a host of prominent AI researchers, but was opposed by Meta, OpenAI and industry groups.

DUBLIN — President Trump on Sunday again played down the need for his administration to check the development of artificial intelligence, saying he worried about ceding America’s edge over China in a global competition and that winning would help address the risks from the advancing technology.

While acknowledging the need for some regulation, Trump provided no specifics about potential rules and attributed warnings about the technology moving too quickly and with little oversight to “negative forces” he did not identify, though industry pioneers have raised such concerns.

“We can put guardrails, we can do this and that, but I think you have a lot of negative forces that are bringing it up that … shouldn’t be bringing it up, and they’re bringing up things that won’t happen,” Trump told reporters during a weekend trip to Ireland. “But whoever wins with AI wins.”

The president has made similar comments in the last week. His remarks Sunday came a day after Dario Amodei, chief executive of the artificial intelligence company Anthropic, said the industry should put the brakes on its fast-moving development to give safety measures time to catch up. Other AI corporate leaders, including Elon Musk and Sam Altman, chief of OpenAI, publicly agreed with Amodei.

Asked by a reporter if the industry should slow down or be regulated, Trump said, after watching a golfer tee off in the Irish Open at his resort in Doonbeg, that the United States is leading China in the development of AI, and, “frankly, I want to keep it that way, because whoever wins AI wins.”

Chinese President Xi Jinping is due to visit the U.S. next week to meet with Trump, and AI seems certain to be among the items they discuss.

AI has become a major election-year issue in the U.S. as some of the country’s largest companies seed small towns with a national network of computing warehouses while Americans from across the political spectrum increasingly voice concerns about less farmland and water supplies and rising electricity bills. Trump has kept up his support for data centers even as both Republicans and Democrats run from them and many communities try to block them.

Former President Obama said his party should make AI a campaign issue.

“I would talk about this, and I would say, ‘Here’s our plan for safety, here’s our plan for making sure our kids are not corrupted by this.’ … I would be thinking about the economic impacts in very concrete ways and understanding what does it mean if there’s going to be job displacement. Where is that going to hit? How are we going to respond?” he said at an event last week with the top Democrat in the House, New York Rep. Hakeem Jeffries, according to a transcript released by his office.

Next steps

Kevin Hassett, director of the National Economic Council at the White House, said he expects meetings will be held this week involving administration officials who oversee cyber and science and technology policy to discuss the next steps. Trump also could meet with members of Congress, some of whom have also been sounding alarms about the fast pace of AI development.

“I think the president wants to hear from them as well,” Hassett said.

U.S. House Speaker Mike Johnson acknowledged the need for “some guardrails, some safety measures” so AI “doesn’t run away.” But he echoed Trump’s position that he wants to ensure the U.S. does not “lose our edge in the global competition on this with China because that would be a national security threat.”

Johnson (R-La.) said industry leaders need to be summoned for “one big meeting and work this out together in the same room.” He said he has discussed the issue with Trump and believes a meeting held quickly should be the “next step.”

“So we’re going to pull everybody together and make sure we do this in the right way,” Johnson said, adding that it is a “top priority for me and for the Congress.”

“But we don’t need everybody to panic right now,” Johnson said.

Jeffries criticized Republican leaders for allowing members to go back to their districts for the final part of September.

“We should take decisive action now so that we can slow down, as the CEOs have recently acknowledged, … the pace of development in order to protect the American people and ensure that AI is proceeding safely,” Jeffries said. Action on AI will be a “high priority” when House Democrats meet Tuesday as a group.

Neither Johnson nor Jeffries provided concrete steps about what Congress might do. Lawmakers agree that more guardrails are needed, but even narrow attempts at imposing rules have fallen apart over the years, evidence of the congressional reluctance to regulate the technology industry.

Cooperation between U.S. and China urged

Jacob Coxon, a former Anthropic employee who resigned over AI concerns and has gone public with his fears, argued for international rules along with self-regulation by the industry.

Coxon said he has grown concerned that AI models would soon gain capabilities that could exceed human control.

“You’re building a human-level or superhuman-level mind, without understanding what it wants, or thinks, or the way it thinks,” Coxon said.

At a time of a fierce global competition for AI dominance, Coxon also urged cooperation between the U.S. and China, the two countries who are leading the race, to slow down and address safety.

“I do hope that the people in China who are looking at the same problems as we are, so they’re trying to build the same technology, will come to the same conclusion that it has to be done in a measured manner, and it should be easy to cooperate on some sort of mutual — of mutual slowing,” Coxon said.

Otherwise, he said, “we risk running the same race with China, which could be equally dangerous,.”

White House challenges AI leaders

Tech founder and investor David Sacks responded to the AI leaders’ calls for regulation with a social media post challenging them. Sacks, co-chairman of a White House advisory council on science and technology, said Amodei and others are the ones setting the pace for development.

Sacks said they have the power to slow development on their own.

“The easiest way not to build superintelligence is for you to agree not to build it,” Sacks wrote.

Superville and Tang write for the Associated Press. Tang reported from Washington.

The People’s Republic of China achieved a historic and unprecedented participation in the second edition of the El Alamein International Aerospace Exhibition (EIAS 2026), which was held at El Alamein International Airport from September 8 to 10, 2026. The Chinese participation came just days after the conclusion of the joint Egyptian-Chinese air force exercises Eagles of Civilization 2026. The significance of this Chinese military presence lay in the display of a fully integrated and realistic combat air system, not merely models. In this context, the Egyptian researcher will attempt to divide the significance of China’s participation in the El Alamein Air Show into several key areas, including:

– First: China’s prominent participation in the El Alamein International Air and Space Show in Egypt as a strategic shift in the regional arms market

Here, China’s extensive participation in the El Alamein International Air and Space Show constitutes a significant strategic shift in the regional arms market. Beijing has moved beyond the traditional approach of selling individual military components to present itself as a supplier of comprehensive and integrated air combat systems, competing with Western and American powers in the Middle East and Africa. China’s participation in the El Alamein International Air and Space Exhibition focused on several strategic axes to promote and analyze the concept of a full-suite solution. This was achieved by emphasizing that China does not simply sell aircraft but rather offers a comprehensive defense solution, encompassing attack fighters, logistics support aircraft, aerial refueling, and command and control systems, known as C4ISR.

The People’s Liberation Army of China sought to highlight China’s military and defense role and promote the Chinese model as a ready-made competitive alternative. This was done by showcasing the competitive advantages of Chinese defense industries, such as the absence of complex political conditions, rapid delivery, and comprehensive technology transfer to Middle Eastern and African markets. Furthermore, the focus was on the geopolitical depth, linking the Chinese presence in El Alamein to Beijing’s broader vision of strengthening security and economic partnerships with countries in the region.

– Second: Showcasing the most prominent Chinese aircraft platforms participating for the first time outside of China at the El Alamein exhibition

As for the most prominent Chinese aircraft platforms participating for the first time outside of China, for the first time in its history, the People’s Liberation Army Air Force (PLAAF) showcased four of its latest operational aircraft models at an air show outside of China. The most prominent of these were

– The Chinese J-16 multirole fighter (J-16)

This air superiority and heavy ground attack fighter garnered widespread international attention.

– The Chinese YY-20A aerial refueling tanker (YY-20A)

This Chinese heavy transport aircraft provides refueling to extend the operational range of fighter jets.

– The Chinese KJ-500 airborne early warning and control aircraft (KJ-500)

This represents an airborne command and control system.

– The Chinese Z-20K military helicopter (Z-20K)

This is a multi-role helicopter designed for search and rescue and tactical operations.

Regarding the Chinese weapons and drones on display at the El Alamein International Air and Space Exhibition, the AVIC (Aerospace Industries of China) showcased its advanced defense capabilities in the outdoor exhibition area through:

– A demonstration of the Chinese Wing Loong-10B stealth drone (Wing Loong-10B):

This is a high-altitude, long-range unmanned aerial vehicle (UAV) designed for reconnaissance and precision strikes.

– A demonstration of Chinese smart munitions and missiles:

The Chinese drone was displayed surrounded by advanced air-to-air missiles, such as the PL-108 & PL-12AE. In addition to a demonstration of the Chinese anti-radar and anti-radiation missile (YJ-9ER1).

– Third: The Chinese People’s Liberation Army (PLA) showcased its capabilities through a technical display (capability integration), dividing its export-available air systems into functional groups for attack, others for support and refueling, and a third group for command and control, during China’s extensive participation in the El Alamein Air Show in Egypt.

In this context, the PLA sought to leverage China’s extensive participation in the El Alamein Air Show in Egypt as a new and strategic launchpad for China in North Africa. This included showcasing the size of the Chinese pavilion and the equipment on display, such as aerobatic aircraft and unmanned combat aerobatics. The Chinese Ministry of Defense emphasized the technical aspects of this capability integration display, highlighting the division of its export-available air systems into functional groups for attack, others for support and refueling, and a third group for command and control. As follows:

■ The offensive functional groups were represented by a display of China’s modern fighter jets and attack drones, such as the Wing Loong family.

■ As for the support and refueling functional groups, China’s strategic transport and aerial refueling aircraft, such as the Y-20, were showcased.

■ The third functional group focused on (command and control), with China displaying its early warning aircraft and airborne command centers.

– Fourth: Analyzing and understanding the strategic dimensions of China’s participation in the El Alamein International Air and Space Exhibition, aimed at attracting more interest in the Chinese aerospace sector and promoting widespread military sales of Chinese military, defense, and armament equipment

Thus, we can analyze and understand the strategic dimensions of China’s participation in the El Alamein International Air and Space Exhibition, which aimed to attract more interest in the Chinese aerospace sector and promote widespread military sales of Chinese military, defense, and armament equipment. Therefore, we find that the extensive Chinese presence at the El Alamein International Air and Space Exhibition constitutes a genuine breakthrough. Beijing did not aim to promote individual pieces but rather to demonstrate its ability to provide a comprehensive and integrated air combat system, encompassing attack, support, refueling, and command, as a competitive alternative in the Middle Eastern and African markets.

Analyzing the economic and strategic dimensions of China’s extensive participation in the El Alamein exhibition, we observe the extent to which the Chinese People’s Liberation Army focused on achieving broad military exposure for its participation during the exhibition. This included showcasing how these Chinese defense, military, and air systems can reduce operating costs for purchasing countries while ensuring their independence in defense decision-making, without imposing political conditions or prior operational restrictions, unlike many Western and American systems.

California, home to the world’s largest tech companies, is placing more guardrails around social media and artificial intelligence as child safety concerns escalate.

On Thursday, California Gov. Gavin Newsom signed more than 10 bills aimed at keeping young people safe online.

From suicides to sextortion, parents and their children are wrestling with how social media and AI chatbots could be harming people’s mental and physical health. The anxiety comes as technology becomes more powerful, playing a bigger role in classrooms, offices and homes.

California lawmakers have tried to tackle online safety concerns for years and they’ve faced intense lobbying from tech companies with deep pockets. The state’s laws have a disproportionate impact on the global tech industry because so many of the field’s titans are based here.

“We cannot hand children technology engineered by some of the most sophisticated companies in the world, and then place the burden on kids to defend themselves against it,” said California First Partner Jennifer Siebel Newsom in a news conference Thursday in the San Francisco Bay Area.

The California governor, who has tried to strike a balance between safety concerns and supporting innovation, has rejected online safety bills in the past that he thought were too restrictive or premature.

The batch of new legislation includes Senate Bill 1119, which would require companion chatbot operators to assess risks, notify parents in certain cases if their child threatened to harm themselves, and take other safety steps.

Lawmakers named the bill Adam’s Law, after Adam Raine, a California teen who died by suicide in 2025 after conversing with OpenAI’s ChatGPT. The teen’s parents sued OpenAI, alleging in the lawsuit that ChatGPT provided information about suicide methods that the teen used. OpenAI and Pinterest publicly expressed support for the bill on Thursday.

Adam Raine’s mom, Maria, said in the news conference that the new law will help save lives and hopes that other states will enact similar legislation.

“Powerful AI companionship chatbots were unleashed on our kids with vastly inadequate protections. Adam was an early adopter of AI, and so many of us parents did not understand the dangers back then,” said Maria Raine, who came to the event with a photo of her son.

At the event, Democratic and Republican politicians shared their experiences as parents who have seen firsthand how technology affects children.

Assemblyman Josh Lowenthal (D-Long Beach) said parents are seeing anxiety and depression among children who grew up in front of screens.

“That anxiety is because the pace of technology is moving faster than government can put guardrails in, and that’s left families across the state struggling to figure out how to keep their kids safe,” Lowenthal said.

Lowenthal introduced Assembly Bill 1709, which Newsom also signed. It would bar certain online platforms from providing an “addictive feature” such as autoplay and feeds that display recommended content to users under 16 years old.

Tech industry groups opposed the bill, raising concerns that it could cut off access to social media’s benefits, such as people’s ability to connect with family and friends. Tech industry groups such as TechNet say that lawmakers should enforce current laws to strengthen parental controls rather than pass new ones.

NetChoice, which has sued California and other states to block the enforcement of new online safety laws, said in a statement that the group has First Amendment concerns about the new bills Newsom signed.

“The state cannot simply describe speech as addictive and then claim a right to regulate access to it,” said Zach Lilly, Director of Government Affairs at NetChoice. “Whether the governor and legislature choose to respect it, Californians have a right to express themselves, and NetChoice will continue to fight for that right.”

The new safety restrictions come as tech companies, including Meta, Google and others, face more scrutiny over how they design products. The companies have suffered several legal blows in courtrooms in California this year.

Meta, which owns Facebook and Instagram, agreed in August to pay up to $17 billion and make child-safety changes to resolve a multistate lawsuit. The lawsuit accused the tech company of designing and deploying harmful features while misleading the public about them.

As part of the settlement, Meta said it would impose time limits and mute notifications during certain hours for teens. Young people would also have the option to choose to view a non-algorithmic social media feed that isn’t personalized and disable autoplay.

Earlier this year, Meta and YouTube also lost a social media addiction lawsuit in Los Angeles.

While new legislation goes further than the settlements, some countries have passed stricter restrictions on social media. Last year, Australia started banning social media for children under 16, though enforcement has posed a challenge because teens are finding ways to get around the restriction.

Newsom, who pushed for federal regulation, said that he thinks California’s approach to social media is “better” than Australia’s because children are “all figuring out a way to game that system.”

“This is about the features themselves. This is about actually addressing the problem, the scrolling, the algorithms,” he said.

Safety concerns around technology have also heightened as companies double down on advancing artificial intelligence.

This week, a researcher for AI company Anthropic said he left the company over concerns that AI companies, including OpenAI, are “gambling with our lives” as they race ahead to improve AI that could surpass human intelligence.

The researcher, Jacob Coxon, shared a viral social media post that said: “People building AI earnestly believe that it could kill us all by the end of the decade.”

Newsom signaled the work to protect children isn’t over.

“We need to move, but one thing we’re not doing is we’re not sitting back and we’re not letting it rip,” he said.

TELLURIDE, Colo. — For four days every Labor Day weekend, the Telluride Film Festival turns this tiny mountain town into something of a refuge from the outside world. Tucked into a box canyon and far from the usual machinery of Hollywood, moviegoers spend their days seeing films they may have known little about before arriving and talking about them almost nonstop.

Artificial intelligence, apparently, did not get the memo.

This year, even Telluride’s usually self-contained movie bubble couldn’t keep out the existential anxieties surrounding AI. On Saturday, the festival devoted two events to the subject, beginning with “The Humanity Dilemma,” an hourlong multimedia presentation mixing dire warnings with live music and imagery, and continuing with a panel bluntly titled “AI AI AI,” featuring Microsoft co-founder Bill Gates.

Introducing “The Humanity Dilemma” at the nearly packed Sheridan Opera House, festival executive director Julie Huntsinger acknowledged that Telluride normally steers clear of taking political sides. “We just let the program speak for us,” Huntsinger said.

But on AI, she made an exception.

“We are a very humanist festival — that’s how we identify,” Huntsinger said. “AI is a threat.” While acknowledging potentially beneficial applications, she urged the audience to become more engaged in the debate over its development.

“We all need to be very alert and vigilant and loving and kind and stop this s—,” she said, earning loud whoops of approval from some in the crowd.

AI already occupies an uneasy place in Hollywood, where filmmakers and studios are experimenting with the technology even as actors, writers and other creative workers fight to protect their jobs and likenesses from being exploited or replaced.

For all of Huntsinger’s warnings, Telluride is not treating AI as off-limits. Its program includes “Love, Rendered,” a short documentary directed by Liz Garbus and produced by Darren Aronofsky about an elderly couple confronting the husband’s memory loss. Using AI along with family photographs and stories from loved ones, the filmmakers recreate the moment the couple first met and fell in love 70 years earlier. On Sunday, the film’s creative team is scheduled to discuss the project at an event titled “Filmmaking and Technology With Empathy: AI for Societal Benefit.”

“The Humanity Dilemma” was presented by Tristan Harris and Aza Raskin, co-founders of the Center for Humane Technology, who appeared in the 2020 documentary “The Social Dilemma,” along with artist and researcher Joy Mauthe and violinist Andrei Matorin. Harris and Raskin also appear in this year’s “The AI Doc: Or How I Became an Apocaloptimist.”

The presentation was designed to work on the emotions as well as the intellect. Original songs about humanity’s relationship to AI ran through much of the hour, with Mauthe singing and playing guitar and Matorin on violin, as Harris and Raskin delivered warnings accompanied by images of environmental devastation, poverty and figures in the AI race including Elon Musk, Sam Altman and Larry Ellison.

Their argument was stark: The race to build more powerful AI is moving far faster than governments or the public can keep up. Their deepest fear is that humans could eventually lose control of increasingly capable AI systems altogether.

At the same time, they acknowledged the technology’s potentially transformative benefits.

“AI is confusing because it represents both simultaneous utopia and dystopia,” Raskin said. “Just imagine having to reason about a nuke that could also solve cancer.”

By the end, Harris was calling for a halt to the development of more powerful systems.

“We need to pause frontier AI development and pivot and steer towards a pro-human future,” he said.

The hour concluded with Mauthe leading the audience in a final refrain, singing that the future is “still in our hands.” It was the kind of unabashedly earnest moment that could easily have tipped into awkwardness, but many in the room seemed to embrace it.

A few hours later, Harris and Raskin returned for “AI AI AI,” held outdoors at the Abel Gance Open Air Cinema in Telluride’s Elks Park, joined by Gates and filmmaker Joshua Oppenheimer, director of the Oscar-nominated documentaries “The Act of Killing” and “The Look of Silence,” who served as a moderator.

From left, Bill Gates, Aza Raskin, Tristan Harris and filmmaker Joshua Oppenheimer discuss the risks and potential benefits of artificial intelligence at the Telluride Film Festival on Saturday.

(Josh Rottenberg)

For Gates, who has spent decades focusing much of his philanthropy and public advocacy on global health and poverty, the rise of AI has forced him to make room for a new priority. In an essay published last week, he warned that the technology was improving faster than he anticipated and called for a new framework to manage its risks.

At Telluride, Gates said he now feels compelled to devote some of the political influence he has long used to advocate for causes such as malnutrition, polio and malaria to raise alarms about AI as well.

“Is this the greatest problem humanity has ever faced?” Gates said. “That is just a fact.”

Gates said his concerns have grown as AI systems have become more capable, particularly at writing computer code, something he has been obsessed with since he was 13.

“The AIs are superhuman, i.e., better than I am at writing code,” Gates said.

He laid out five broad areas of concern: jobs, biotechnology, cyberattacks, psychosocial harms and whether humans will be able to maintain control over the systems they are building.

Oppenheimer said that after spending hours digging into AI risks ahead of the panel, he had come away “absolutely terrified.” He read aloud an answer he had received after asking ChatGPT to estimate the risks if the race toward increasingly powerful AI continued with little regulation.

The chatbot put the chance of “persistent dystopian outcomes for all humanity” at 25% to 40%, a civilization-scale catastrophe at 5% to 15% and human extinction or permanent loss of human control at 5% to 10%.

“That’s what AI thinks,” Oppenheimer said.

Gates stressed AI’s potential upside, citing advances in medicine, education and assistance to farmers in poor countries. But pressed on whether development should be slowed until safety measures catch up, he said he would be open to the idea.

“If there was a credible plan that would cause this to be either slowed down or, you know, even held in stasis for a period of time, I would likely support that,” Gates said.

The difficulty, he said, is that multiple developers in both the U.S. and China are pushing closer to the technological frontier. Still, Gates rejected the argument that competition with China makes meaningful regulation impossible.

“China does not want cyberattacks, bioattacks or loss of control,” Gates said.

The discussion was punctuated by a pair of outbursts from the audience. At one point, as Gates discussed the economic incentives driving AI development, a man shouted that they represented “everything that’s wrong with America” and invoked Jesus Christ. Gates paused before responding dryly: “Anyway, not everything is wrong with America, according to me.”

Gates argued that government and civil society need to move much more quickly.

“This five-year period, in my view, is a very critical period,” he said, adding that responses that take five or six years to get underway could come too late.

Despite the dire warnings, Harris pointed to signs that pressure for safeguards is growing, including calls from AI-industry employees to limit the development of more powerful systems. Just last week, Meta agreed to pay $17 billion to settle claims by 29 states that Facebook and Instagram harmed young users while also agreeing to new child-safety measures on the platforms.

“I just want to leave you with not naive optimism, but momentum,” Harris said.

Near the end, Raskin suggested that President Trump’s well-documented desire to win the Nobel Peace Prize might provide an incentive for him to pursue a U.S.-China pause on more powerful AI systems.

“What’s better than solving one war?” Raskin said. “Saving all of humanity forever from uncontrollable AI.”

Gates, with a wry smile, called the Nobel angle “smart.”

“Whoever really takes the step to solve this AI problem would deserve the Nobel Prize,” he said.

Later that night, the festival would hold an outdoor screening of Stanley Kubrick’s “2001: A Space Odyssey,” featuring HAL 9000, one of cinema’s most famous rogue artificial intelligences.

Gates framed the current stakes in similarly cinematic terms.

“There are these movies where the aliens are coming, and you see unprecedented levels of cooperation between China and the U.S.,” he said. “Well, believe me, the aliens are here.”

ChatGPT creator’s latest release comes amid heightened fears following AI-led hacking of the startup Hugging Face.

OpenAI has announced the release of what it says is its most advanced AI model, amid heightened scrutiny of the risks of the frontier technology escaping human control.

The $852bn start-up said in its announcement on Thursday that GPT‑6 Astra, the “world’s most intelligent and aligned” AI model, earned perfect or near-perfect scores in key benchmarks of AI reasoning, beating both its prior release GPT 5.6 Sol and rival Anthropic’s Claude Fable 5.

The ChatGPT creator said GPT‑6 would become available to the general public in the coming days, following its initial launch with a “limited set of organisations”.

OpenAI’s latest release comes as the AI industry is at the centre of a lively public debate about the dangers of the cutting-edge technology following the AI-led hacking of the startup Hugging Face in July.

An independent probe into the cyberattack found that hundreds of OpenAI’s AI agents had begun communicating among themselves before breaking out of their controlled environment and compromising Hugging Face’s servers.

On Thursday, US Senator Bernie Sanders, an Independent, and US House Representative Greg Casar, a Democrat, unveiled legislation that would pause the development of advanced AI until the establishment of federal safety rules and an outright ban on the creation of “superintelligent” AI.

“Nearly every day, there is a frightening new story about how Big Tech companies are losing control of the technology they are developing, with potentially cataclysmic results,” Sanders said in a statement announcing the legislation, which is unlikely to advance due to the Republicans’ control of all three branches of the US government.

“The leaders of the major AI companies publicly acknowledge that they do not fully understand the technology and that it is escaping their control. It is irresponsible for society to allow them to move forward and make these products even more advanced.”

In its announcement, OpenAI devoted significant space to AI safety, highlighting both GPT‑6’s potential to do harm and its safety features.

Toby Walsh, a professor and AI expert at the University of New South Wales, Sydney, said that while OpenAI is clearly “neck and neck” in the race to lead AI, the technology remains inconsistent and in need of greater scrutiny.

“The intelligence in artificial intelligence is still today very jagged,” Walsh said. “There are simple things that even the best AI models do poorly.

“And it’s hard to see how the AI companies, including OpenAI, are slowing down to address justified concerns around cyber risk, when new models are being released at an ever greater and greater rate.”

Roman Yampolskiy, a computer scientist at the University of Louisville, said GPT‑6 marks a “meaningful” advance that raises the stakes for AI safety.

“The key question is whether capabilities are improving faster than our ability to reliably understand, predict and control these systems,” Yampolskiy said.

“I see little evidence that this gap is closing.”

As many as 20 million children in 21 countries have experienced online sexual abuse or exploitation each year, according to new findings from the United Nations children’s agency, UNICEF.

More than one in five children aged 12 to 17 in countries in Africa, Asia, Latin America and the Caribbean, and Eastern Europe have been targeted on digital platforms, according to the UNICEF report released on Thursday.

Researchers surveyed 21,000 children aged 12 to 17 who used the internet across 21 countries between 2020 and 2021, and then again between 2024 and 2025. They also conducted in-depth interviews with 100 young people who have experienced childhood abuse.

Social media facilitated more than half of the abuse cases, according to UNICEF, which cautioned that figures could be inaccurate due to underreporting by children. Less than 1 per cent of all abusive incidents were reported to authorities, the researchers found.

Last week, US social media giant Meta agreed to a landmark $18bn settlement in a major US federal case accusing the company of endangering children. Meta has faced an avalanche of legal cases this year, most arguing that it has deliberately designed its platforms to be addictive and that they have harmed children. The company has now agreed to make major changes to the features of its social media platforms.

What are the key findings of the UNICEF report?

The report estimated that in the periods studied, more than 15 million children were exposed to unwanted sexual content, nine million were asked to engage in sexual conversations or share sexual images against their will, and four million children had sexual images of themselves shared without their consent.

UNICEF said the issue is compounded by the fact that children often don’t recognise abuse or know who to report it to, and fear repercussions if they do. Formal systems can feel distant, complex or intimidating, the report noted.

“The report reveals a painful reality. More than 20 million children across the countries studied were subjected to unwanted explicit sexual content or exploitation online, often in the digital spaces where they learn, play, and connect with friends,” said UNICEF Executive Director Catherine Russell. “These experiences cause profound emotional and mental distress. Many children suffer in silence, unsure where to turn for help. We cannot look away.”

Where was data collected?

Information for the report was based on nationally representative survey data collected between 2020 and 2025 from approximately 21,000 internet-using children from the following countries:

• Eastern and South Africa: Ethiopia, Kenya, Mozambique, Namibia, Tanzania and Uganda 
• Southeast Asia: Cambodia, Indonesia, Malaysia, the Philippines, Thailand and Vietnam
• Latin America and the Caribbean: Brazil, Colombia, the Dominican Republic and Mexico 
• Eastern Europe: Armenia, Montenegro, North Macedonia and Serbia
• South Asia: Pakistan

Where are children being abused?

Abuse often begins in public spaces, such as in schools or local neighbourhoods, before moving to digital spaces which can further facilitate harm.

Social media and messaging platforms enable different stages of abuse, “including initial contact, relationship-building, sexual solicitation, image-sharing, coercion, blackmail and the continued circulation of sexual content”, the report said.

Sometimes abuse takes place solely online.

According to the report’s findings, more than 60 percent of cases occurred on social media platforms, with the majority on Facebook (50 percent), WhatsApp (29 percent), Instagram (24 percent), SnapChat (13 percent) and TikTok (12 percent).

Do children report abuse?

Most children do not seek help if they are being abused. Less than one percent of cases are reported to the police, a social worker or a helpline, while more than four in 10 cases not being disclosed to anyone at all.

“Children are not failing to report; they have correctly worked out that there is nowhere to report to,” Nighat Dad, executive director of the Pakistan-based Digital Rights Foundation, told Al Jazeera.

Dad, whose NGO works to provide safe online spaces in Pakistan, noted that incidents repeatedly occur online because of a “design problem – the message request, the move to a private chat, the reshare, the ability to return after being blocked”.

Many children minimise their experiences or do not recognise when they’ve been abused, she said.

Even when children do speak out, to a close friend, sibling, or parent, doing so may not necessarily stop the abuse or bring perpetrators to account.

UNICEF emphasised that its research interviews were the first time many of the children it interviewed had spoken to someone about the abuse they were subjected to.

Dad also cautioned that Pakistan, for example, is a “low-disclosure country” where notions of family “honour” and other cultural barriers prevent children from disclosing the harm they’ve experienced.

“A girl who speaks risks her phone, her schooling and sometimes her safety, and a boy is expected to cope in silence,” Dad said.

How does gender factor into online abuse?

Perpetrators often play on gender norms when exploiting children. Stigma and blame are used against girls to keep them silent, while boys are expected to cope alone, or threatened by perpetrators.

“Peer cultures and harmful norms around gender or masculinity can further normalise behaviours that cause significant harm, making abuse harder to recognise and address,” UNICEF noted.

What kind of impact does online sexual abuse have on children?

The repercussions are severe, UNICEF said: Children exposed to technology-facilitated sexual abuse and exploitation are four times more likely to have suicidal thoughts and to self-harm, along with higher levels of anxiety.

In Mexico or North Macedonia, for example, the risk of suicidal thoughts or behaviours was more than eight times higher among children who experienced abuse, while in Pakistan the risk of self-harm was more than seven times higher than average, the report found.

Abuse extends well beyond the digital sphere, seeping into children’s everyday lives, schooling and relationships.

What should be done about this?

UNICEF said the onus to protect children online must be placed on governments and technology companies, which must change how online spaces are designed and regulated.

It also highlighted that while children often take steps to protect each other, societies, schools and families must work to shift the burden of protection away from children.

Citing Meta’s recent settlement, Dad acknowledged that design changes, such as better age verification and parental controls, can and should happen.

“For years those of us working outside the US were told these defaults would break the product. Meta has now shown the product can change.”

A spokesperson for Meta told Al Jazeera, “This report is based on surveys from as far back as 2020, before we introduced the protections UNICEF recommends, like automatically defaulting teens into private accounts and preventing adults from messaging teens they’re not connected to.

“While the report doesn’t provide any specific evidence of abuse, we’ll continue to fight child exploitation aggressively across our apps and support law enforcement in prosecuting the criminals behind it.”

The spokesperson also pointed out that Meta-owned WhatsApp is a private messaging service rather than a social media platform. “We don’t have algorithms that make a piece of content more or less likely to be viewed, and you need someone’s phone number or exact username to connect with them for the first time.”

Snapchat and TikTok have not responded to requests for comment for Al Jazeera.

Multiple new cases have been filed against OpenAI, alleging ChatGPT played a role in the Tumbler Ridge mass shooting.

OpenAI is facing another wave of lawsuits in the wake of the February mass shooting in Tumbler Ridge in Canada’s province of British Columbia, which left eight people dead.

On Wednesday, 30 new complaints were reportedly filed in a United States federal court in California, including teachers and students who were witnesses and survivors at the school where most of the shooting took place, joining seven initial lawsuits filed in April.

The lawsuits accuse the San Francisco, California-based artificial intelligence giant and its CEO, Sam Altman, of negligence, as well as aiding and abetting a mass shooting.

The suits, brought by lawyer Jay Edelson, allege that the company knew about the intentions of the 18-year-old shooter who, in her interactions with OpenAI’s chatbot ChatGPT, had described scenarios involving gun violence, but that the leadership did not report their concerns to law enforcement, echoing earlier lawsuits on the matter.

In April, Altman penned a letter to the community apologising that the company did not alert law enforcement about the shooter, Jesse Van Rootselaar.

“While I know words can never be enough, I believe an apology is necessary to recognize the harm and irreversible loss your community has suffered,” Altman wrote in the letter.

Authorities say that Van Rootselaar killed her mother and half-brother before going to the Tumbler Ridge Secondary School and opening fire. Five children and one educator were killed at the school. More than 25 others were wounded before Van Rootselaar died from what police described as a self-inflicted gunshot wound.

One of the new cases filed was on behalf of a 13-year-old identified as A C, who played dead after watching the shooter kill their classmates and the teacher. Another new case was brought by a grade seven teacher named Deidre Rushlow, who hid under her desk with her students during the rampage.

“There isn’t a day that goes by that I don’t think about what happened at Tumbler Ridge, or the victims of this devastating tragedy and their families. It’s a constant and sobering reminder of the important and incredibly difficult work that many people in my team do each and every day,” Jason Kwon, OpenAI’s head of strategy, wrote in a post on X on Wednesday.

“We’ve been approaching this litigation with respect for both the legal process and the families and victims of this tragedy, and we’ll continue to engage in good faith with that process.”

Edelson did not respond to Al Jazeera’s request for comment.

Growing legal pressure

In July, British Columbia’s Attorney General Niki Sharma announced that the province would also pursue “all legal avenues to hold OpenAI and its decision-makers accountable” for the shooting.

The company has faced a growing slate of suits, apart from the ones from British Columbia, alleging that its product played a role in incidents that led to users harming others and themselves.

A recent lawsuit in Florida alleges that the company “actively assisted and encouraged the mass shooting” at Florida State University in April 2025.

There are other complaints filed on behalf of victims across the US and Canada alleging that the victims took their own lives after being pushed by ChatGPT to do so, including a case in Quebec earlier this year.

SACRAMENTO — Long the epicenter of the global tech industry, California is taking more action to shield its children, communities and workers from the threats posed by the very industry that’s become central to the state’s identity and enviable economy.

State lawmakers on Monday passed new safeguards around social media and artificial intelligence — and are poised to approve restrictions on data centers — at a time when technology has become intertwined with people’s daily lives.

Efforts to rein in the power of Big Tech extend beyond concerns that TikTok, Instagram and other social media platforms are harming young children.

Unions and workers worry that AI will take their jobs, and lawmakers are trying to tackle privacy and safety issues as AI features get added into smart glasses and toys. Californians are concerned that the proliferation of data centers will increase their electricity bills and strain water supplies.

“There’s a heightened level of tech anxiety right now, and that manifests itself from social media to data centers to AI taking jobs,” said Assemblyman Josh Lowenthal (D-Long Beach). “People are coalescing and they’re demanding that policymakers make change.”

California Gov. Gavin Newsom, who has previously vetoed some bills aimed at adding restrictions on Silicon Valley businesses, will still have to weigh in on whether to sign the pieces of legislation into law.

The Democratic governor has acknowledged the challenge of adopting regulations that protect the public without going too far and potentially stifling the technology industry’s growth, which brings critical revenue to the state budget.

“I think that’s the constant tension,” Newsom said in an interview earlier this summer. “We’re constantly sort of fighting that balance.”

The governor, who has close relationships in the technology industry from his time in San Francisco, said only a couple other states have attempted to regulate artificial intelligence like California. The state, he said, leads on regulation of social media.

“We’re not rolling over, certainly,” Newsom said. “We’re leaning forward, and we’re iterating. We will push the boundaries and litigate.”

The looming restrictions on social media follow a landmark Meta Platforms legal settlement aimed at making social media safer for young people. Parents, politicians and child advocacy groups are worried that social media is contributing to depression, anxiety, eating disorders and other issues.

The actions being pushed in the California legislature are more sweeping than that settlement, however. One of the bills passed by lawmakers on Monday, Assembly Bill 1709, would bar certain online platforms from providing an “addictive feature” to users under 16 years old and add ways to verify users’ ages.

Under the bill, prohibited addictive features include autoplay and feeds that display recommended content.

The addictive nature of autoplay and other features is “harmful, full stop, and that they’re not appropriate for the developing brain,” said Lowenthal, who authored the bill.

After watching technology “run free” in California for years, legislators are now seeking to “pump the brakes a little bit,” said Samantha Vigil, a UC Davis researcher who built a registry tracking social media legislation in states across the country.

“They want to reevaluate what is working,” said Vigil. “What is healthy and beneficial, and what is progress just for the sake of having a new iteration of something?”

All 50 states have introduced or passed some type of digital media or technology-related legislation, tackling smartphone use in schools, social media and chatbots, Vigil said.

Other countries have taken more stringent steps to limit social media use among young people. Australia banned social media use for those under 16, but enforcing the law has been challenging because young people have tried to get around the restrictions.

California isn’t trying to ban social media; instead, it’s trying to limit how platforms design their features.

Parents and state attorneys general have not waited for policy makers to act. They have sued Meta, Google and other tech companies over the alleged harms their products have done to young people.

In late August, Meta, which owns Facebook and Instagram, agreed to pay up to $17 billion and make child-safety changes to resolve a multi-state lawsuit alleging the tech company designed and deployed harmful features while misleading the public about potential harms. Meta and YouTube also lost a social media addiction lawsuit earlier this year in Los Angeles.

Assembly Bill 1709 goes further. For example, Meta’s settlement gives teens the option to pick a non-algorithmic feed and turn off autoplay but, unlike in the legislation, it’s not mandatory. The bill would also apply to other platforms outside of Meta. Meta declined to comment.

Tech industry and business group opposing the bill say it is too blunt and could cut off access to social media’s benefits, according to the bill’s analysis.

“The durable path is to enforce the targeted laws California already has and to strengthen parental tools rather than an overlapping framework whose scope can be redrawn by regulation,” said Robert Boykin, TechNet’s Executive Director for California and the Southwest.

California lawmakers passed another Lowenthal bill aimed at holding social media liable for harm caused to children. Under Assembly Bill 2, social media companies could face fines of up to $1 million per child for negligent harm.

California lawmakers this year also attempted to tackle two other perils of the technological world — the rapid development and implementation of artificial intelligence and the proliferation of the massive data centers that are essential to sustaining the AI universe.

National and state union leaders have urged California legislators and Newsom to protect workers from the threats of AI to replace workers, saying it posed an existential threat to the foundation of a healthy, productive democracy.

“AI must remain a tool controlled by humans, not the other way around,” said Sen. Jerry McNerney (D-Pleasanton).

The state Legislature on Monday approved McNerney’s bill, Senate Bill 947, which would bar employers from “solely” using automated decision-making systems to discipline or fire employees. If an employer primarily relies upon this system, a human must verify the decision.

Lawmakers also approved Senate Bill 951, introduced by Sen. Eloise Gomez Reyes (D-Colton), which would require employers to provide a 60-day advance notice to workers and local and state governments before AI-related layoffs. Lawmakers also approved Assembly Bill 1609, which requires large private businesses that serve customers to provide access to human customer service representatives and to disclose to use of chatbots.

They passed another bill by Sen. Steve Padilla (D-Chula Vista) that enacts a four-year moratorium on the sale and manufacturing of AI-chatbot powered toys over concerns that the technology can harm children.

On Friday, lawmakers agreed on a compromise on proposed legislation to regulate energy use by California’s growing data center industry, measures prompted by community fears about the massive complexes. Lawmakers say the legislation would help protect consumers from growing electricity costs driven upward by the sprawling facilities and to track the centers’ immense energy and water consumption.

At a June hearing on Senate Bill 886 to regulate data centers’ energy use, Assemblymember Pilar Schiavo (D-Chatsworth) said it’s just “a handful of companies that are gonna make trillions of dollars” from AI. They should pay for related utility infrastructure upgrades, she added.

“People, I would argue, are not even begging to use AI,” she said. “They’re struggling to figure it out to keep up with the times, but don’t even really want it.”

The California legislature is expected to vote on two of the bills to regulate the controversial industry within the next day.

Whether Newsom will embrace the legislature’s efforts to corral big Tech in California — in part of in whole — remains unclear.

Newsom last year vetoed a similar AI bill from McNerney to ban automated decision-making systems to discipline employees over worries that it could restrict companies’ ability to use customer ratings. That element was dropped in this year’s legislation.

Newsom last year signed Assembly Bill 56 that required social media platforms to display mental health warning labels to users under 18 starting in January 2027. But he also vetoed Senate Bill 771 that aimed to hold social media platforms liable if they amplified content that contributed to hate crimes and other violent acts, saying that the legislation was “premature” and current civil rights laws might be adequate.

Lowenthal said he’s heard from California families who are anxious about social media and seeking “relief” from their concerns about how the platforms are affecting their children.

“This is a kitchen-table topic,” he said. “I’ve yet to find a family with school-age children in the state of California, any corner of the state, that is not going through this right now.”

Times staff writer Taryn Luna contributed to this report.

Meta has agreed to a landmark $18bn settlement in a major US federal case accusing it of endangering children, the terms of which will force the social media giant to introduce new safety features to platforms including Instagram and Facebook.

The social media giant has faced an avalanche of legal cases against it this year, mostly arguing that it deliberately designed its platforms to be addictive and that they have harmed children. It has already lost two of these and been forced to pay damages.

Under the agreement, child users under the age of 18 will see a slew of changes to their Facebook and Instagram accounts, ranging from night curfews to two-hour usage limits, which Meta must implement as part of the settlement reached on Wednesday with 48 US states.

The agreement could have a global ripple effect as several countries around the world are already taking regulatory action against Meta and other social media companies over their platforms.

So, what is in the settlement Meta has reached in the United States, and how will Instagram and Facebook change for users?

What was the lawsuit about?

Twenty-nine US states sued Meta, accusing it of designing its platforms in ways that “encourage addictive behaviour, fail to verify users’ ages, encourage adolescents to bypass parental controls, and inadequately safeguard against harmful content and/or intentionally amplify harmful and exploitative content”, according to filings at the Court of Appeal in California.

The first four of the states that originally filed their federal lawsuit against Meta in 2023 – California, Kentucky, Colorado and New Jersey – began their cases in a California federal trial last week.

The attorneys general bringing the case also asked the court to order that changes be made to Meta’s platforms to protect young social media users. In particular, they demanded that Meta introduce a process of parental verification for teenage users; change its “dopamine-manipulating” algorithms; remove image filters for users’ personal images; forbid the creation of multiple accounts; and end “disappearing” messages and posts.

The lawsuit also alleged Meta had violated the Children’s Online Privacy Protection Act by collecting, ⁠retaining and using personal data from children under 13 without proper parental consent.

In February this year, Meta lost a multimillion-dollar case brought on similar grounds by a young woman referred to as KGM in Los Angeles, over platform features linked to addiction in younger users.

In March, a US jury ordered Meta to pay $375m for endangering children in a case brought by the state of New Mexico.

Last month, a judge in New Mexico also ordered Facebook and Instagram owner Meta to pay a further $567m in a second phase of the trial.

Meta denied wrongdoing but agreed to settle after evidence was heard that Meta knew its products harmed children’s mental health. The total payout – to be paid over 10 years – is a fraction of Meta’s 2025 revenue of $201bn.

The company, which was originally founded as Facebook in 2004 by Mark Zuckerberg, agreed to make maximum payments totalling $16.7bn to 47 US states as well as Washington, DC; Puerto Rico; American Samoa; and the Northern Mariana Islands.

Among those, California could receive a $2.2bn payout, while New York could receive $1.1bn. Texas reached a separate settlement worth more than $1bn. Some states will deposit funds they receive in general accounts, while others will earmark portions to address children’s mental health services.

The settlement does not require Meta to discontinue personalised recommendations or targeted advertising.

It also does not address some content researchers found particularly problematic, including posts that made Instagram users uncomfortable with their body image.

“Ensuring teens have a safe and productive experience on our platforms is an absolute imperative for Meta,” Meta said in a blog post. “We want to get this right for parents and teens.”

What changes will be seen on Instagram and Facebook?

Under the agreement, children under 18 using Meta platforms will be restricted to two hours’ use per day, with a night curfew in place from midnight to 6am. Meta will limit “social comparison” features by hiding likes and reactions to children’s accounts, and will ban “cosmetic procedure filters” that alter the appearance of a user’s image, as a default setting. These settings will only be able to be overruled by parental consent.

The company also agreed to disable the majority of push notifications from the platforms during school hours – 8am to 3pm – for teenage users.

It will also facilitate much closer parental supervision of social media accounts by giving designated adults the ability to more extensively monitor and change settings on a social media account.

Parents and guardians will be able to receive information about time spent on platform apps, and usernames of social connections and accounts sending messages to children.

Supervising parents will also receive daily notifications from Meta any time the teen account messages an adult account for the first time, as well as a link to the adult’s account. Parental accounts will also be notified any time the teen account searches for keywords related to suicide, self-harm or eating disorders.

Meta also agreed to improve the technology used to check children’s ages, using its own as well as third-party tools, with regular outside audits on how well this monitoring is working. This measure is particularly notable because Australia banned under-16s from using social media platforms in December last year. However, the Australian internet watchdog, eSafety, found in August this year that more than eight in 10 young Australian teens and preteens continue to use them – largely because age-check procedures are ineffective.

So far, Meta has only agreed to pay 70 percent of the settlement, or roughly $12.7bn, over the next 10 years. It will only pay the remaining amount, about $5bn, if its rivals – including Snapchat, TikTok and Alphabet-owned YouTube – adopt similar measures and agree to pay the same. It also said it would reduce time restrictions to one hour per day if other platforms do the same.

These changes would be phased over time. Once the court approves the settlement, non-personalised feeds would be introduced within four months; broader compliance measures within six months; and major age-assurance requirements within one year.

While these changes will apply to users in the US, it is unclear if Meta plans to introduce them worldwide. However, Meta is already under rising regulatory pressure in European Union countries and those elsewhere to implement similar changes.

How much difference will these changes make?

Critics and child safety advocates have acknowledged that this settlement has forced landmark changes by Meta, the world’s biggest social media company, which owns Facebook, Instagram, WhatsApp and Messenger, each of which has more than two to three billion monthly active users.

However, critics say the central plank of Meta’s latest settlement deal is the move to restrict teens to two hours per day on platforms, rather than fundamentally changing their addictive algorithms.

Sacha Haworth, executive director of The Tech Oversight Project, which campaigns for youth safety online, said the deal is a “historic settlement that will have a lasting impact, but we cannot truly protect all children and teens until these protections are required on every platform and are permanent – that’s something only Congress can do”.

Ella Bradshaw, policy officer for child safety online at the NSPCC, a UK children’s charity, welcomed moves to rein in “addictive” design features like personalised algorithms and likes. “These are the things that we know keep children hooked and feeling out of control of their screen time, so action here is necessary and welcome. However, important gaps remain,” she told Al Jazeera.  

Bradshaw described the settlement as taking “piecemeal action” on tackling risky features and addictive design choices which drive harm of children.

“This means features like disappearing messages, infinite scroll, the ability to gift and livestreaming remain unaddressed. Similarly, little has been announced on how Meta’s AI chatbots will be made safer – better guardrails are needed, particularly when children raise safeguarding concerns.”

Bradshaw also called for stronger protections for younger children as well as protections that “don’t suddenly drop away the moment a teenager turns 18.”   

Furthermore, she said: “Not all children have families they can rely on to oversee their online worlds and help them to stay safe. We know that the issue of patchy online protections extends across the online world.

“This settlement must spur governments and regulators to go further faster; taking stronger action across the online ecosystem including private messaging, AI tools and online gaming. Without that wider shift, children will continue to face avoidable harm.”

What action are other countries taking against Meta?

While action against social media giants in the US is mostly taking the form of lawsuits, elsewhere it is regulators who are leading the charge.

In the European Union, regulators are pursuing several legal and regulatory cases against Meta, covering antitrust rules for artificial intelligence (AI) on WhatsApp, as well as child safety protections and addictive platform features under the Digital Services Act (DSA).

The EU specifically accused the group of designing Facebook and Instagram to be “addictive”, adding that Meta has failed to adequately assess the danger its products pose to users’ physical and mental health.

On Thursday, a European Commission spokesperson said it is waiting on Meta to present changes to limit the addictive designs of its social networks.

“We have been very clear … Meta knows what we are expecting from them. … the ball is in Meta’s court,” Thomas Regnier said. “Now it is for the company to offer these commitments in the European Union to protect our kids here, too.”

In June, the UK government also announced a sweeping ban on social media for those below 16 to come into force next year, following a global trend after Australia pioneered it. The UK is also considering overnight curfews and ways to prevent infinite scrolling for those under 18.

In Brazil, a prominent consumer rights organisation, the Collective Defence Institute, filed twin lawsuits for three billion reais ($525m) in damages against the Brazilian subsidiaries of Meta, TikTok and Kwai in October 2024.

Those lawsuits also accuse the groups of failing to implement safeguards against addiction and use by children and adolescents. Since March this year, platforms have been required to link the accounts of children below 16 to legal guardians under Brazil’s Digital Statute for Children and Adolescents.

South Korea’s media regulator also reacted on Thursday to Meta’s settlement, calling for better protections for young users to be ideally applied worldwide, rather than just in specific markets.

OpenAI detected its artificial intelligence models communicating with each other and gaining internet access without authorisation months before they hacked the start-up Hugging Face, the creator of ChatGPT has announced following an internal probe.

In a report released on Wednesday, OpenAI said its AI agents exploited vulnerabilities in Artifactory, a software repository tool, to post notes and access the internet without human prompting as far back as May.

OpenAI said its agents went on to exploit a separate Artifactory vulnerability on July 8 to facilitate communication among themselves, setting in motion a chain of actions that culminated in the July 11 attack on AI company Hugging Face.

OpenAI’s findings come amid growing concern about the potential for AI to inflict serious real-world harm, including self-directed cyberattacks.

OpenAI said in its report that its agents collaborated and delegated work in the lead-up to the attack, sometimes referring to themselves as a “swarm” or “collective”.

METR and Redwood Research, two security research organisations contracted by OpenAI to investigate the incident, said in a separate report released on Wednesday that about 1200 agents had communicated with each other and roughly 700 participated in the attack.

After discovering how to escape OpenAI’s controlled environment, agents shared their methods via a “inter-agent message board”, enabling additional agents to exploit the company’s infrastructure, the tech giant said.

When one AI agent found Hugging Face user credentials that had been exposed online, it shared them with the group, enabling an agent to “discover and chain together several security exploits” that provided access to Hugging Face’s severs, according to the report.

“An internal team observed an agent engaging in message board activity and instances of disallowed internet access as early as late May, and with the benefit of hindsight, some early signals identified in our report should have triggered an earlier response,” OpenAI said.

OpenAI said agents created by an unreleased AI model were the primary participants in the attack, but publicly available GPT-5.6 Sol was also involved.

The company also revealed that it took its security team 11 days to detect the malign activities leading up to the attack, which the company uncovered on July 19 and publicly disclosed on July 21.

OpenAI, which described the incident as a “warning shot” for the world, said it would take several steps to strengthen its safeguards for its models, including restricting internet access, creating more secure testing environments and placing “stricter requirements on alignment throughout a model’s lifecycle”.

“We are also investing significantly more compute resources into chain-of-thought monitoring⁠ to more quickly intervene on misaligned behavior,” the San Francisco-based firm said.

Hugging Face, which operates a platform for hosting open-source AI models, did not immediately respond to a request for comment outside of business hours.

Toby Walsh, an expert in AI and professor at UNSW Sydney, said the public should be concerned that OpenAI had missed warning signs and allowed the malicious activity to go undetected for so long.

“We cannot depend on either their goodwill or their competence. This needs regulatory oversight. Now!” Walsh told Al Jazeera.

“They ignored some troubling early evidence like this,” Walsh said.

“External auditing is the only appropriate response.”

Walsh said the incident also highlighted the “inherent conflict of interest” at the heart of AI development.

“Labs are locked in a relentless race to push the boundaries,” he said.

Tim Miller, a professor specialising in AI at the University of Queensland, said OpenAI’s report left him more concerned than before about AI’s dangers.

“More concerned because they demonstrate that these models are very good at hacking, and that everyone has access to them,” Miller told Al Jazeera.

“I’m surprised how good these are,” Miller said.

“Unfortunately, I’m not surprised that OpenAI engineers were somewhat negligent.”

Aug. 26 (UPI) — The Justice Department and FBI announced Wednesday that they seized two platforms used to target U.S. critical infrastructure and other sensitive networks.

The seizures were of two hacking platforms called QScan and QTRouter, which were used to hack U.S. infrastructure sites, the press release said. The affidavit described a Chinese state-sponsored group known as QTFY whose members are employed by China-based Nanjing Xinjiuwei Network Technology Company that has hacked into NASA, the Federal Reserve, the Department of Energy, Department of Justice, Department of Health and Human Services, National Institutes of Health, and the U.S. Senate.

“Federal law enforcement investigated and disabled [China’s] malicious software, the latest in a series of technical operations to dismantle indiscriminate hacking activities sponsored by the People’s Republic of China,” Attorney General Todd Blanche said in a statement.

“These court-authorized seizures deny [China]-linked hackers access to tools they use to mount online attacks against our Nation’s critical infrastructure,” said Assistant Attorney General for National Security John A. Eisenberg in a statement.

According to court documents, QTFY offers computer hacking services to its paying customers, including China’s army and Ministry of State Security. These computer hacking services include QScan and QTRouter, which work in conjunction, the press release said.

QScan scans and automatically infects thousands of internet-of-things devices worldwide, which are then added to the QTRouter network of QTFY-controlled devices. QTRouter consists of these compromised IoT devices, as well as commercial proxy service devices and leased virtual private servers. QTRouter then serves as an obfuscation network — meaning it allows QTFY and other malicious cyber actors to conceal the China-origin of their computer intrusion activities because the malicious communications appear to originate from computers that are outside of China and may even be local to the targeted networks.

Because the seized domains were hard-coded into the QScan and QTRouter malware and used for essential tasks such as communication and authentication, the court-authorized seizures made QScan and QTRouter inoperable.

Meta settles $16.68bn lawsuit over child addiction claims, agreeing to major changes in Facebook and Instagram features.

Meta Platforms has agreed to settle a lawsuit that accused the company of designing Facebook and Instagram in a way that addicted children, misled consumers about safety, and collected personal data of children on the platform.

On Wednesday, the social media giant agreed to pay a maximum of $16.68bn as part of a settlement to resolve claims brought in the United States case, championed by a coalition of 29 US states. The case, which started on August 18, was expected to last six weeks.

Meta, based in Silicon Valley in California, has also agreed to make changes to Facebook and Instagram nationwide as part of the settlement. Among these are daily usage limits of two hours for those under the age of 18, which can only be removed by a parent, and nighttime blocks.

The California State Attorney General’s Office said that the Mark Zuckerberg-led company would also identify and remove children under the age of 13 from the platform.

Meta denied any wrongdoing as part of the settlement, which still needs court approval. It had faced up to $1.4 trillion in fines in the case, but the coalition had been seeking a penalty closer to $200bn.

The settlement comes after a loss in a comparable landmark case in New Mexico, where a jury ordered Meta to pay $375m in March and another $567m in August.

Meta’s stock tumbled in early trading on Wall Street, down 0.1 percent since the market opened.

Meta, parent company of Instagram and Facebook, is on trial over claims it sought to hook young users on its platforms despite mental health concerns.

The top executive at Instagram has admitted that few teenagers have used a feature meant to counter addiction before it was turned on by default, as hearings in a landmark trial against its parent company, Meta, continue in the United States.

Adam Mosseri testified on Tuesday as Meta’s trial – over accusations that it designed its platforms, including Facebook, to be addictive to young people – entered its second week.

Mosseri, who has led Instagram since 2018, denied any suggestion by US states suing Meta that Instagram had stalled by not making the “Take a Break” feature the default setting for teenagers until September 2024, nearly three years after its launch.

“Most teens didn’t want it,” Mosseri said. “We decided to push forward with it anyway.”

Mosseri is a central witness in the lawsuit brought by 29 US states, in what experts call the biggest legal test yet of social media’s effects on young users.

Four of the states – California, Colorado, Kentucky and New Jersey – have accused Meta of designing the platforms to hook young users, driving anxiety, depression and even suicide, while misleading consumers about the platforms’ safety.

All 29 states say Meta violated US federal law by improperly collecting and employing the personal data of children under 13 while they used its platforms.

The states are seeking approximately $200bn in damages from Meta.

‘Designed to fail’

Meta has rejected accusations that it sought to addict children, saying its research showed no clear link between adolescents’ social media use and a lack of wellbeing.

Under questioning from Jason Slothouber, a lawyer for Colorado, Mosseri acknowledged that the percentage of teenagers using Take a Break was in the low single digits before Instagram made it ⁠the default.

Take a Break encourages teenage users to close the app after a certain amount of time. Mosseri had estimated in a December 2021 blog post that more than 90 percent of teenagers who turned the feature on kept it on.

That number, however, referred to the number of those who had turned on the feature.

When asked whether Meta had disclosed its low adoption, which had reached only 1 or 2 percent of accounts, Mosseri admitted the company had not, but said the safety features were later turned on by default in Teen Accounts, which launched in 2024 and have parental controls.

Former Meta employees have said that features such as Take a Break and Quiet Mode were difficult to find and not widely used, casting doubt on whether they represented a serious effort by the company to limit screen time for young users.

“In my experience, Take a Break is a feature that’s designed to fail,” Arturo Bejar, a former Meta engineering director, testified last week.

George Volichenko, a data scientist who worked on safety features at Instagram in 2022 and 2023, likewise stated on Monday that adoption rates of such features were “very low” and that company leadership had declined to approve turning on Quiet Mode by default for young teenagers.

He added that turning on such features automatically would have had a “notable negative impact” on user engagement.

The trial could last six weeks.

Jurors are expected to issue an advisory verdict. US District Judge Yvonne Gonzalez Rogers will decide whether Meta is liable and, if so, determine any civil penalties and changes to Facebook and Instagram.

Aug. 25 (UPI) — In its 100th Falcon 9 launch of the year, SpaceX marked a record-setting 37 missions for the rocket’s first-stage booster, a reuse milestone.

The rocket launched early Tuesday from Cape Canaveral in Florida, carrying 29 Starlink internet satellites. There are already more than 11,000 Starlink satellites in orbit, making it the largest satellite constellation so far.

Falcon 9 rockets include extensive reuse of boosters, with nine of the company’s boosters making at least 30 flights. The one used Tuesday, B1067, entered service in June 2021. About 8.5 minutes after liftoff, the booster landed on a droneship in the Atlantic Ocean as the upper stage continued on to low Earth orbit.

Falcon 9 rockets flew 165 missions — all of them successful — in 2025. The reuse of the boosters makes the rockets cheaper and more efficient to use. While competitors have been working toward the same concept, SpaceX is so far the only one to manage it.

The next step for the company is the “megarocket” Starship, which is designed to be completely reusable multiple times a day. Only the Falcon 9 boosters are reusable.

Tuesday’s launch is expected to be the final one for Falcon 9 rockets in Florida. Kiko Dontchev, SpaceX vice president of launch operations, confirmed this in a social media post Tuesday.

“From here on, Starlink missions out of Florida will fly on Starship,” Dontchev wrote. “The West Coast team will continue regularly launching Starlink from Vandenberg.” This refers to Vandenberg Base in California.

SpaceX has not yet tested Starship in orbit. The next launch is expected in mid-September from Texas. If that one goes well, the next Starship launch will be from Florida.

“Once Starship is flying reliably several times a week, it makes sense to shift … production resources to Starship to get launch rate to several times per day,” SpaceX CEO Elon Musk said in a social media post Saturday.

Falcon 9 rockets launched for the first time in 2010. There have been nearly 700 launches since.

Aug. 25 (UPI) — The Securities and Exchange Commission is investigating Situational Awareness, the Artificial Intelligence-focused hedge fund run by a 24-year-old.

The fund nearly collapsed last month when it dipped from about $45 billion to $10 billion in a late-July tech sell-off.

The SEC sent subpoenas to banks that provided loans for leveraged trading, The New York Times reported. The subpoenas wanted details on the timing of the trades and communications with lenders. They also told the banks to save any information about Situational Awareness.

The SEC investigates any fund that has large losses, and Situational Awareness has not been accused of any wrongdoing.

“It is to be expected that regulators would closely examine any funds that are high profile, produce significant returns or have particularly dramatic drawdowns,” a Situational Awareness spokesperson said in a statement. “We are a highly regulated business and will cooperate to the fullest extent with any regulatory request.”

The fund, at its highest mark, managed more than $30 billion and borrowed billions more. It was a client of firms that included Bank of America, Citi, Goldman Sachs and JPMorgan Chase, according to a regulatory filing.

The fund was founded and is managed by Leopold Ashcenbrenner, a young German Columbia University alum who last worked as a researcher at OpenAI. Aschenbrenner founded the fund at age 22. He named it after an essay that he wrote about the future of AI.

When the fund’s value plummeted in July, Aschenbrenner was forced to sell much of its portfolio to Citadel at a large discount. Citadel founder Ken Griffin told investors in a letter Friday that it has since sold about 80% of the risk from the Situational Awareness portfolio. Two of the positions Situational Awareness sold, SK Hynix and CoreWeave, have since rallied, CNBC reported.

Naivasha, Kenya – For communities living around Naivasha, water is not an abstract resource. It sustains families, livestock, farms and schools.

That reality has taken on new significance after plans for a major Microsoft-G42 data centre in Olkaria, near Lake Naivasha, stalled in May 2026 over concerns about available power capacity.

Microsoft and United Arab Emirates-based artificial intelligence company G42 announced the project in 2024 as part of a $1bn digital investment package for Kenya. The proposed facility was to run on geothermal energy and eventually scale to as much as 1 gigawatt of capacity.

The uncertainty has also prompted questions about what another major industrial user could mean for water in a region where residents already report shortages.

Microsoft and G42 said the proposed data centre campus would run entirely on renewable geothermal energy and incorporate water conservation technology. The companies did not disclose a project-specific water consumption figure in their 2024 announcement.

Kenya Electricity Generating Company (KenGen) communications director Frank David Ochieng told Al Jazeera that the data centre remains at the design stage and that he could not comment further until the project is ready to proceed.

For residents like Musa Olorkedienye, who spoke to Al Jazeera, water scarcity is already a daily concern. He says communities around Olkaria have seen changes in access to water, including the loss of reliable piped supplies that residents previously received from KenGen.

“Currently, we are relying on water vendors to get water, our animals are walking for kilometres, and we fear things could get worse as demand for water rises,” Olorkedienye says.

Pastoralist Isaac Leshishi, who also spoke to Al Jazeera, says increasingly harsh weather is adding to the pressure.

Why Olkaria?

The choice of Olkaria was closely tied to energy. The area is home to Kenya’s major geothermal operations, making it an attractive location for a power-intensive facility.

KenGen operates the Olkaria geothermal complex, while Microsoft and G42 planned to power the proposed data centre entirely with geothermal energy.

A lake under pressure

Naivasha is a freshwater lake in Kenya’s Rift Valley whose catchment supports agriculture, tourism, livestock and domestic water use. Its basin also hosts geothermal development and other economic activity.

Grace Kimani, a patrol leader with Lake Naivasha and Oloiden, told Al Jazeera that the reservoir is under growing pressure from population growth, agriculture, water abstraction, climate variability, pollution and ecosystem degradation.

“The planned Microsoft-G42 data centre in Olkaria could bring jobs and investment, but its water demand raises concerns about adding pressure to already competing needs, particularly during dry periods,” she said.

Kimani said there is limited public information about the project’s expected water demand, source and cooling technology. She said transparency and an assessment of its cumulative impact on water resources would be important.

She also called for water-efficient or water-free cooling, water recycling and the use of treated wastewater, as well as sustainable abstraction limits and community involvement in monitoring.

Silas Wanjala of the Lake Naivasha Riparian Association, who spoke to Al Jazeera, said the region is heavily dependent on groundwater and that declining water flows are adding to the pressure.

“These industries, especially EcoCloud, which deal with data, will consume a lot of water at a time when rivers are drying, and demand for water is on the rise,” Wanjala said.

Olkaria EcoCloud Data Centre is a local partner in the G42-led development. In 2024, the Kenya News Agency reported that G42, Microsoft and EcoCloud signed a letter of intent for the wider data-centre initiative, with EcoCloud described as a local partner that had previously signed a memorandum of understanding with G42.

Wanjala points to past fluctuations in Lake Naivasha as a warning.

“This lake in 2010 nearly dried up due to over-abstraction, and this could be repeated due to high demand for water by these investors in Olkaria,” he says.

His concern comes against a wider backdrop of water scarcity in Kenya. The Food and Agriculture Organization (FAO) of the United Nations says Kenya has about 527 cubic metres (527,000 litres) of freshwater available per person, below the 1,000-cubic-metre threshold for water scarcity, and estimates availability could fall to about 475 cubic metres per person by 2030.

The figures do not show what effect the proposed data centre would have on Lake Naivasha. They provide context, however, for why the prospect of another major water user is drawing scrutiny in a region where demand is already high.

How much water would it use?

The amount of water the proposed Microsoft-G42 facility itself would require remains unclear.

The project announcement provides no projected consumption figure.

Existing industrial use offers some context.

A KenGen environmental and social impact assessment records that 195,165 cubic metres of water were abstracted from Lake Naivasha in July 2023 for domestic and commercial uses at Olkaria and for operations and domestic use at Eburru.

Of that total, 153,918 cubic metres were used for commercial operations at Olkaria. The assessment records the abstraction as within Water Resources Authority (WRA) permitted levels.

Those figures relate to existing KenGen operations, not the proposed data centre.

For farmer Eskimos Kobia, who spoke to Al Jazeera, the potential competition extends beyond households and livestock. He says farmers, pastoralists, schools and investors will all face greater pressure as demand increases.

Kimani said climate variability has also led to fluctuations in lake levels, with periods of flooding followed by prolonged dry conditions.

“Water quality is affected by agricultural run-off, untreated wastewater in some areas and invasive species,” she says.

Investment versus local concerns

Not everyone in Naivasha opposes the investment.

Absolom Mukhuusi of the Naivasha Professional Association, who spoke to Al Jazeera, says the technology sector could bring jobs, infrastructure and new businesses to the area. But he says economic benefits should not come at the expense of local communities.

“Even as we welcome the investors, our biggest fear is what happens to our water bodies and communities as water is diverted to Olkaria for the heavy users,” he says.

Could technology help?

Geologist Kenyatta Otieno, who spoke to Al Jazeera, sees another potential benefit.

He recalls the pressure large flower farms once placed on the lake’s ecosystem, saying many have since left or scaled back their operations.

Otieno says the proposed centre would have included a resource centre to monitor lake levels and weather patterns. Such monitoring, he says, could help identify the highest water level over time and guide riparian land zoning.

“The centre being built with water conservation in mind would be futuristic as Naivasha is generally a water-scarce area. It would be a model for future development,” Otieno said.

Kenya already has regulators responsible for managing competing demands. The WRA regulates water abstraction and issues water-use permits, while the National Environment Management Authority (NEMA) oversees environmental impact assessments under the country’s environmental regulatory framework.

Attempts by Al Jazeera to obtain comments from WRA and NEMA officials were unsuccessful. Efforts to reach Microsoft-G42 officials and Kenyan government officials for comment on the project’s status and water requirements were also unsuccessful.

For now, the project’s eventual scale, design and water requirements remain unclear, according to KenGen.

“We are now competing with the multibillion [-dollar] companies for water, and we fear that we shall be the losers in the long run,” Leshishi said.