Artificial Intelligence

In an era of escalating climate disasters, the ability to translate data into life-saving action has never been more critical. For the Asia-Pacific region—the world’s most disaster-prone, this is not an abstract challenge but a daily reality. At the forefront of this battle is the United Nations Economic and Social Commission for Asia and the Pacific (ESCAP), which is leveraging artificial intelligence to close the gap between risk knowledge and on-the-ground resilience. In this exclusive Q&A, Kareff May Rafisura, Economic Affairs Officer at the ICT and Disaster Risk Reduction Division of ESCAP, provides a clear-eyed look at their innovative tool, SatGPT, and how it’s changing the game for communities from the remote village to the ministerial office.

1. It’s one thing to see a flood risk map, and another to break ground on a new levee. Could you walk us through how a local official might use SatGPT to confidently decide where to actually build?

Kareff May Rafisura, Economic Affairs Officer at the ICT and Disaster Risk Reduction Division of ESCAP: First, it’s worth noting that there’s growing rethinking within the science and policy communities on the long-term benefits and trade-offs of constructing artificial levees.

Going back to your question, understanding an area’s flood history is key to making smart infrastructure decisions. You wouldn’t build a levee on natural floodplains, for example. Without risk knowledge, levees might not protect communities effectively and could even cause problems downstream or in ecologically sensitive areas. SatGPT offers a rapid mapping service that helps local officials make risk-informed decisions. It significantly reduces the time and cost traditionally required to assess flood characteristics, such as frequency, spatial extent, and impacts, and converts that data into actionable information. This information is critical for decisionmakers who must weigh it alongside economic, social, and environmental considerations when determining whether, and where, to build a levee.

2. We often hear about getting tech “to the last mile.” Picture a rural community leader with a simple smartphone. How does SatGPT’s insight practically reach and help them make a life-saving decision?

Kareff: SatGPT’s strength lies in enhancing historical risk knowledge. It’s not designed to predict the next disaster, but rather to help communities prepare more effectively for it. For instance, when a rural leader needs to decide whether to evacuate ahead of a flood, she will still rely on early warnings from national meteorological services. What SatGPT can do is support smarter ex-ante planning—so that when early warning information arrives, the community is ready to respond quickly. This includes decisions on where to build shelters, how to lay out evacuation routes, and where to preposition relief supplies. These are all critical elements that must be in place to help avert disasters, as consistently demonstrated in the cyclone response histories of India and Bangladesh.

3. Floods are an urgent threat, but what about slower crises like droughts? Is the vision for SatGPT to eventually help with these less visible, but equally devastating, disasters?

Kareff: ESCAP coordinates the long-standing Regional Drought Mechanism, which has been supporting drought-prone countries in gaining access to satellite data, products, tools, and technical expertise—everything they need to conduct drought monitoring and impact assessments more effectively. Our support goes beyond making data available—we work with countries and partners to strengthen institutions and capacities, converting these data into actionable analytics and insights. We are currently working with three Central Asian countries in establishing their own Earth observation-based agricultural drought monitoring systems.

4. AI is powerful, but it can sometimes reflect our own blind spots. How are you ensuring SatGPT doesn’t accidentally worsen inequality by overlooking the most vulnerable communities in its models?

Kareff: You raised a valid concern. That’s why in our capacity development work, our participants combine SatGPT’s flood mapping with socio-economic data to pinpoint who’s most at risk and where. They work on use cases that unpack the exposure of essential services like hospitals and water treatment facilities. When these critical infrastructures fail, it’s the poorest who pay the highest price. That’s why it’s vital to understand the hazards that threaten them.

5. Governments have tight budgets. If you were making the pitch to a Finance Minister, what’s the most compelling argument for investing in SatGPT now versus spending on recovery later?

Kareff: Investing in reducing disaster risk – which involves measures taken before disasters occur to reduce vulnerability and enhance resilience (e.g., early warning systems, resilient infrastructure, land-use planning) – is far more cost-effective than recovery. Every dollar invested in disaster risk reduction can save multiple dollars in future losses. While the benefits are context-specific, a recent multi-country study found that for every $1 invested, the return can be as high as $10.50.

6. The region is innovating fast, with countries like Indonesia and Thailand building their own systems. How does SatGPT aim to be a good teammate and connect with these national efforts, rather than just adding another tool to the pile?

Kareff: That’s a good point. And beyond technological innovation, we’re also seeing progress in policy and institutional innovations being put in place. Our intention is not to replace national systems, but to show what’s possible when you make risk knowledge accessible and actionable. We work closely with our national counterparts with a focus on integrating SatGPT insights into existing workflows and systems-not reinventing them.

7. Training young professionals is key. Beyond the technical skills, what’s the most important lesson you hope they take away about using this technology responsibly?

Kareff: I’m glad you recognize that today’s most pressing need goes beyond technical expertise. That’s precisely why our technical capacity-building activities are held alongside youth forums to provide a platform for young people to engage in meaningful conversations around values and motivations. As stakeholders, we all share the responsibility of upholding safe, secure, and trustworthy artificial intelligence systems to support sustainable development.

8. Looking ahead a year, what would a “win” for SatGPT look like on the ground? Is it a specific number of communities better protected, or a faster warning time?

Kareff: Forecasting and enhancing the forecast lead times remains the responsibility of mandated early warning agencies. SatGPT is well-positioned to support efforts to protect more communities. By enhancing the historical understanding of floods, it can help improve the accuracy of early warning information, help communities proactively plan their response, and reduce disaster risk ex-ante. In that sense, I would say that effective SatGPT roll-out would amount to both gains in space and time – more communities being warned with improved lead times for mitigative response with more reliable historical data for granular risk characterization.

9. The document mentions turning the Jakarta Declaration into action. From your vantage point, what’s the biggest spark of progress you’ve seen so far?

Kareff: One of the most promising sparks of progress has been the strengthened regional cooperation aimed at enhancing the capacity of countries—especially the countries in special situations—to overcome barriers to accessing the benefits of innovative geospatial applications. With the support of ESCAP members, we are implementing field projects, providing capacity-building and technical assistance, facilitating expert exchange, and knowledge sharing across more than a dozen countries. These efforts are helping to develop space-based solutions from the ground up to tackle sustainable development challenges such as urban poverty, air pollution, droughts, floods, and crop biodiversity loss.

10. Finally, behind all the data and code, you mention this is about protecting lives. Has working on SatGPT given you a new perspective on what “resilience” truly means for a family facing a flood?

Kareff: Having lived and worked for the United Nations in some of the world’s most flood-prone countries, I’ve witnessed first-hand how the lack of historical data can lead to underinvestment in risk reduction. Tools like SatGPT and other digital innovations are not silver bullets, but they help close this gap by converting geospatial data into actionable insights – quickly and more accessibly – to guide communities to prepare and protect lives and livelihoods.

The conversation with Kareff May Rafisura underscores a pivotal shift in disaster risk management: from reactive recovery to intelligent, data-driven preparedness. SatGPT represents more than a technological achievement; it is a practical instrument of empowerment, ensuring that from the finance minister to the rural community leader, the best available knowledge informs the decisions that save lives and safeguard futures. In the fragile balance between human vulnerability and environmental force, such tools are not just helpful, they are essential. The future of resilience in the Asia-Pacific is being written today, not in the aftermath of disaster, but in the proactive, thoughtful application of innovation like SatGPT.

DAVID Bowie and Kate Bush have been named among the top cultural icons who most accurately predicted how we live today, according to research.

A poll of 2,000 adults found George Orwell, Roald Dahl, and even Ross from Friends – who in 1999 predicted AI would be smarter than us by 2030 – made the top 10 list.

Other cultural icons included Captain Kirk – who used to talk to computers, foldable communicators, and tablets as far back as the 1960s – and Ridley Scott.

The director’s seminal 1982 film Blade Runner is still hailed today as a masterclass in technological foresight.

Meanwhile the sitcom, The Simpsons has a history of uncanny predictions, including Donald Trump‘s presidency, the Pandemic, a FIFA scandal, and the development of smartwatches.

The research was commissioned by Samsung for its ‘Visionary Hall of Fame’ and rounding off the top 10 are musicians Prince and Bjork – with the former predicting online dating and virtual relationships in his album 1999, released over 40 years ago.

STARGIRL

Moment Hollywood star shocks with claim she TIME TRAVELLED to see David Bowie

STAR GONE

Iconic Radio DJ known as ‘Jed The Fish’ who worked with Bowie & Sting dies at 69

While Bjork foresaw the rise of social media in the 1990s and 2000s, predicting that technology wouldn’t just be functional, it would become deeply personal.

Fearne Cotton has teamed up with the brand, as part of their Can Your Phone Do This campaign which highlights the capabilities of Galaxy AI, to go back to her chart show roots, in a brand-new countdown video which reveals the visionaries who feature on the list.

The broadcaster and author said: “These ten icons didn’t just dream about the future; they made it a reality. It’s incredible to see that the future they envisioned is already here, right at our fingertips.

The research also found self-driving cars (39 per cent) topped the list of real-world innovations people remember seeing in pop culture before they became a reality.

This was followed by artificial intelligence (39 per cent) and video calling (33 per cent), along with voice assistants (28 per cent) and smart watches (22 per cent).

Those polled were also quizzed on their use of AI apps or assistants, with 24 per cent using these on their phones daily.

Many use them to ask factual questions (43 per cent), compose messages or emails (22 per cent), and edit photos and videos (22 per cent).

For 23 per cent, they are even translating speech or text among the most used AI functions.

In fact, almost seven in ten (68 per cent) also agreed that today’s AI-powered smartphones feel as though you are carrying the future in your pocket.

Annika Bizon, from Samsung, added: “68 per cent of Brits are amazed that these once-futuristic predictions are now part of everyday life, with over half crediting AI for boosting general knowledge and creativity.

“With Galaxy AI, we’re not just keeping pace with the predictions of modern-day visionaries, we’re actively shaping what comes next.

“We’re turning tomorrow’s possibilities into today’s realities, because when you hold the future in your hand, you’re not just ahead of the curve—you’re defining it.”

SACRAMENTO — Though hailed by some for signing new laws to combat antisemitism in California schools, Gov. Gavin Newsom expressed enough reservations about the bills to urge state lawmakers to make some changes.

Supporters of the legislation, Senate Bill 48 and Assembly Bill 715, said it was needed to protect Jewish students on campus, while opponents argued it was broadly written and would stifle free speech and classroom discussions about current events in the Middle East, including the Israel-Hamas war.

Newsom, when he signed the bills, directed legislators to work quickly on a follow-up measure to address “urgent concerns about unintended consequences.”

The governor made similar requests for nearly a dozen other major bills he signed into law this year, including measures providing safeguards on artificial intelligence, protections for children online and banning law enforcement officers donning masks — a direct response to federal agents hiding their identities during immigration raids across the state.

Newsom’s addendums provide a glimpse into the sometimes flawed or incomplete process of crafting new laws, at times hastily at the end of legislative session, requiring flaws or unresolved conflicts to be remedied later.

San Jose State University professor emeritus and political analyst Larry Gerston said governors sometimes go this route when, despite having concerns, they feel the legislation is too urgent to veto.

“I think you are looking at a situation where he thought the issue was sufficiently important and needed to go ahead and get it moving,” he said.

Gerston, however, noted those with a cynical view of politics could argue governors use this tactic as a way to undo or water down legislation that — for various political reasons — they wanted to pass in the moment.

“Depending upon your attitude toward the governor, politics and legislation, [that viewpoint] could be right or wrong,” he said.

One of the authors of the antisemitism bills, Assemblymember Rick Chavez Zbur (D-Los Angeles), said he will put forth another measure next year and continue working with educational organizations and the California Legislative Jewish Caucus to ensure the right balance is struck.

“The assertions that the bill is intended to prevent instruction about controversial topics, including topics related to Israel, is just not accurate,” said Zbur, who introduced AB 715. “We will be making sure that it’s clear that instruction on complicated issues, on controversial issues, that critical education can continue to take place.”

Zbur said he will reexamine a provision requiring the “factual accuracy” of instructional materials.

“One of the things that we’ve agreed to do was focus on making sure that the bill continues to meet its goal, but revisit that factually accurate language to make sure that, for example, you can continue to teach [works of] fiction in the classroom,” he said.

Another new law flagged by Newsom bans local and federal agents from wearing masks or facial coverings during operations.

The governor approved Senate Bill 627 — carried by Sens. Scott Wiener (D-San Francisco) and Jesse Arreguín (D-Berkeley) — last month as a response to the Trump administration’s aggressive immigration raids that are often conducted by masked agents in unmarked cars. Newsom said it was unacceptable for “secret police” to grab people off the streets.

“This bill establishes important transparency and public accountability measures to protect public safety, but it requires follow-up legislation,” Newsom wrote in his signing statement. “Given the importance of the issue, the legislature must craft a bill that prevents unnecessary masking without compromising law enforcement operations.”

Newsom said clarifications about safety gear and additional exemptions for legitimate law enforcement activities were needed.

“I read this bill as permitting the use of motorcycle or other safety helmets, sunglasses, or other standard law enforcement gear not designed or used for the purpose of hiding anyone’s identity, but the follow-up legislation must also remove any uncertainty or ambiguities,” he wrote.

Wiener agreed to revisit the measure.

“I’m committed to working with the Governor’s office to further refine SB 627 early next year to ensure it is as workable as possible for many law enforcement officers working in good faith,” he said.

California is the first state to ban masking for federal law enforcement and the law will likely be challenged in court. The move drew ire from U.S. Department of Homeland Security Secretary Kristi Noem, who called the legislation “despicable” and said forcing officers to reveal their faces increases their risk of being targeted by criminals.

Newsom is also urging legislators to adjust two new tech-related laws from Assemblymember Buffy Wicks (D-Oakland).

Assembly Bill 853, dubbed the California AI Transparency Act, is intended to help people identify content created by artificial intelligence. It requires large online platforms, such as social media sites, to provide accessible provenance data on uploaded content starting in 2027. Provenance data is information about the origin and modification history of online content.

In his signing statement, Newsom called the legislation a “critical step” but said it could interfere with privacy.

“Some stakeholders remain concerned that provisions of the bill, while well-intentioned, present implementation challenges that could lead to unintended consequences, including impairment of user privacy,” he wrote. “I encourage the legislature to enact follow up legislation in 2026, before the law takes effect, to address these technical feasibility issues.”

Assembly Bill 1043 aims to help prevent children from viewing inappropriate content online. It directs operating system providers to allow parents to input their children’s ages when setting up equipment such as laptops or smartphones, and then requires users to be grouped in different age brackets. It gained approval from tech companies including Meta and Google while others raised concerns.

“Streaming services and video game developers contend that this bill’s framework, while well-suited to traditional software applications, does not fit their respective products,” Newsom wrote in his signing statement. “Many of these companies have existing age verification systems in place, addressing complexities such as multi-user accounts shared by a family and user profiles utilized across multiple devices.”

The governor urged lawmakers to address those concerns before the law is set to take effect in 2027.

Wicks was unavailable for comment.

Published on
21/10/2025 – 9:22 GMT+2

ADVERTISEMENT

Kenny Loggins has reacted to Donald Trump using his song ‘Danger Zone’ in the president’s “disgusting” AI-generated video showing himself wearing a crown, flying a “KING TRUMP” fighter jet and bombing a crowd of protesters with feces.

The video was published as a response to the historic “No Kings” protests which took place across the US on Saturday.

The American singer-songwriter recorded the hit song for the soundtrack of the 1986 Tom Cruise movie Top Gun. He has now called for Trump’s video to be taken down on copyright grounds.

In a statement to Variety, Loggins said: “This is an unauthorized use of my performance of ‘Danger Zone.’ Nobody asked me for my permission, which I would have denied, and I request that my recording on this video is removed immediately.”

He continued: “I can’t imagine why anybody would want their music used or associated with something created with the sole purpose of dividing us. Too many people are trying to tear us apart, and we need to find new ways to come together.”

“We’re all Americans, and we’re all patriotic. There is no ‘us and them’ — that’s not who we are, nor is it what we should be. It’s all of us. We’re in this together, and it is my hope that we can embrace music as a way of celebrating and uniting each and every one of us.”

Well put – especially considering the video has provoked widespread outrage online, with many expressing dismay over the way it shows Trump’s clear disdain for people exercising their right to protest.

Social media users accused Trump of having “the maturity and decorum of a 12-year-old boy”, while others commented: “Can’t believe that’s a president of a country.”

Many posts also pointed out that Trump’s “childish” and “disgusting” AI post revealed a transparent representation of his genuine feelings toward the American people. “It tells you everything you need to know about what he thinks about the people of America who are, in fact, America,” one person commented, while another added: “Him taking a dump on the country is the most honest thing he’s ever posted.”

This is far from the first time that Trump and his administration have used artists’ work without authorisation.

There is an extensive list of musicians who have objected to Trump’s authorized use of their songs. These include ABBA, The Rolling Stones, Bruce Springsteen, Rihanna, Neil Young, R.E.M., Woodkid, Beyoncé and Semisonic.

Sinead O’Connor’s estate previously issued Trump with cease-and-desist orders, while Isaac Hayes’ estate sued him for 134 counts of copywright infringement.

Céline Dion also condemned the use of her song from the Oscar-winning film Titanic, ‘My Heart Will Go On’, which was used at one of Trump’s rallies. Dion’s team questioned the song choice, writing: “And really, THAT song?”

Another band which added their name to the ever-growing list of artists who have sued Trump over the illegal use of their songs in campaign videos was The White Stripes. Last year, the rock band highlighted the “flagrant misappropriation” of their hit song ‘Seven Nation Army’. Jack White captioned a copy of the legal complaint in an Instagram post with: “This machine sues fascists.”

The most recent example to date is Metallica, who forced the US government to withdraw a social media video that used their song ‘Enter Sandman’ without authorisation.

This weekend’s “No Kings” protests saw millions of Americans marching against Trump’s administration, opposing the president’s “authoritarian power grab.”

The 18 October protest, the third mass mobilisation since Trump’s return to the White House, drew nearly 7 million people across all 50 states according to organisers. This figure would make it the largest single-day mobilisation against a US president in modern history.

Global stock markets kicked off the week on a strong note after data showed China’s economy performing better than expected despite ongoing trade tensions with the United States. Investor optimism was also buoyed by expectations of Japanese stimulus and a strong outlook for artificial intelligence (AI) companies during the U.S. earnings season.

Why It Matters

China’s stronger-than-forecast GDP growth (1.1% in Q3) and industrial output gains (6.5%) helped calm fears about a global slowdown triggered by U.S.-China trade frictions.
Meanwhile, optimism surrounding AI-driven tech earnings particularly Nvidia continued to lift global equities, reinforcing investor belief in the sector’s long-term profitability.
At the same time, expectations of further U.S. Federal Reserve rate cuts kept global borrowing costs lower and strengthened risk appetite.

Asia: Japan’s Nikkei surged 2.8% to a record high amid hopes of stimulus under likely new Prime Minister Sanae Takaichi.

Europe: The Stoxx 600 rose 0.7% in early trade.

U.S.: Futures pointed to gains of 0.4–0.5% for the S&P 500 and Nasdaq.

Bonds & FX: Treasury yields dipped to 4.02%, while the euro climbed to $1.1662 on a softer dollar.

Commodities: Gold stayed elevated around $4,266/oz, reflecting persistent geopolitical caution, while Brent crude slipped 0.4% to $61.02 on OPEC+ supply signals.

Jason da Silva (Arbuthnot Latham): “There’s still enough scope for healthy returns from big tech; I’m not selling the AI theme yet.”

Kevin Thozet (Carmignac): Warned of “froth” in some AI stocks but said it’s too soon to exit the trade.

Lorenzo Portelli (Amundi): Predicted gold could rise to $5,000 as central banks diversify reserves and the dollar weakens.

What’s Next

Looking ahead, investor attention will pivot to major U.S. corporate earnings that could shape the market’s next moves. Reports from Tesla, Netflix, Procter & Gamble, and Coca-Cola will offer a clearer picture of consumer demand and how well companies are weathering tariffs and inflation pressures. On the policy front, traders expect the Federal Reserve to deliver two more rate cuts by December, a move that could further support equities, weaken the dollar, and sustain global liquidity. However, the upcoming U.S.–China tariff truce deadline on November 10 looms large, and any breakdown in talks could quickly reverse market optimism. Investors will also watch for fresh data on inflation and labor markets to gauge how long central banks can maintain their dovish stance.

With information from Reuters.

Tim Hentschel, CEO of HotelPlanner.com and HotelPlanner.ai, says his virtual travel agents are already taking 50,000 real calls a day and will hit the 100,000 mark before the year is out

“I’m sorry. That was a generic answer. Let me come up with something a little more meaningful.”

There’s something quite novel about a robot apologising to you. Cassandra, one of HotelPlanner.ai’s American representatives, was as quick on the contrition as it was on suggesting hotels in Tupelo, Mississippi, that I might want to stay in.

I chucked a few requirements at Cassandra and it quickly found me a place that ticked all of the boxes. Cassandra even obliged when I asked for a description of the rooms, and then apologised when I cut it off, demanding that those descriptions be a little less lifeless.

Try as I might, my efforts to wind Cassandra up failed. It kept delivering helpful answers and even extra titbits of information without being prompted, all with an upbeat tone of voice against a backdrop of fake call center sounds.

OPINION

Milo Boyd

READ MORE: Brits using AI to plan trips doubles as Gen Z leads use of robo-travel agents

Robo-travel agents such as Cassandra are currently fielding calls on HotelPlanner.ai, where you can choose the language and gender of your call handler before taking it on a test run. It’s working on a beta trial basis in the UK at the moment, but in the US, the world of AI travel agents is very much here already and booming.

Tim Hentschel, CEO of HotelPlanner.com and HotelPlanner.ai, says his virtual travel agents are already taking 50,000 real calls a day and will hit the 100,000 mark before the year is out. He claims that 10% of HotelPlanner.com’s bookings are made by bots.

That’s no small potatoes, given the company’s expected $1.8 billion gross revenues this year and the 1.5 million properties it can book.

Tim is a huge believer in the power of AI to transform customer service in the travel industry and argues that HotelPlanner.com’s agents are trained to be “more helpful and efficient” than humans and already able to “compete head-to-head” when it comes to customer satisfaction and sales.

While Tim says the company has no desire to stop using humans to answer calls, the AI side of the business is growing quickly and could one day field 100% of all calls. Partly because they’ve learned from the best. The large language model bots were trained using eight million human phone calls.

What’s arguably a little unnerving is that customers aren’t told they’re speaking to a robot, and they often don’t realise.

“We only tell customers if they’re speaking to an AI agent if they ask. Sometimes it has come up as customers have asked them out on dates,” Tim told the Mirror.

“We find AI works best with older customers as it’s extremely helpful and patient—sometimes to a fault, as the AI doesn’t understand the money value of time. The question is, if you want a service, who can perform it better? A human or AI with unlimited information?

“The agents have accents. They make an attempt at humour. The robots are programmed always to be helpful, it has nothing but kindness.”

Whether the idea of an AI customer service bot excites or appalls you, their arrival in increasing numbers seems inevitable. Just this week, OpenAI announced it had struck a deal with travel giant Expedia, paving the way for holidaymakers to book trips directly through the platform.

At the Travel and Tourism Summit in Rome at the end of September, Jane Sun, CEO of Trip.com Group, predicted that AI would “double the travel market” – because “people will work three days due to AI, and take much longer holidays.”

Jane went on to describe how she thought AI would improve customer service.

“We must make sure our customers are very well looked after with good customer support. 30 seconds, an AI in your native language will be able to speak to you over the phone. Within 30 seconds a call centre employee will be able to answer your call. Within two minutes of a crisis, our team will be able to call them to get them to safety. This will be powered by AI,” she said.

How do you feel about AI travel agents? Let us know in the comments below or by emailing [email protected].

Some 8% of adults surveyed said they use the technology to give them ideas for where to go on holiday, up from 4% a year earlier, a poll suggests

Double the number of people are turning to artificial intelligence (AI) for holiday planning inspiration compared to last year, fresh research reveals.

Travel industry body Abta, which commissioned the study, branded the technology a “creative co-pilot” that holidaymakers can utilise to research, plan and book their getaways. Around 8% of participants in a survey of 2,001 UK adults carried out in July admitted they use AI to spark ideas for their holiday destinations.

This marks a rise from 4% twelve months ago. Abta’s director of communications Graeme Buck suggested there is “a potential for this acceleration” to persist, stating: “I wouldn’t be surprised if that 8% becomes 16% next year.”

The most recent findings showed that those aged 25-34 are most inclined to harness the technology for holiday brainstorming, with 18% of participants in this age bracket confirming they do so.

OPINION

Milo Boyd

READ MORE: How much money you’ll need in your account to enter each EU country from Sunday

This was followed by 35 to 44 year olds (14%). Among those 65 and over, the proportion dropped to merely 1%.

AI travel applications encompass services including chatbots and resources for translation and itinerary creation.

More than two in five (43%) survey participants indicated they would feel somewhat confident using AI to organise a holiday, though this fell to 38% when it came to actually making reservations.

Steve Heapy, CEO of Jet2, addressed travel industry leaders at Abta’s annual convention in Calvia, Spain, stating that technology like AI will “continue to become a bigger part of what customers do before they book”.

He added: “We’re all going to have to work harder to justify the margin that we want to earn by demonstrating superior knowledge of the product and providing that anticipation and excitement.”

Neil Swanson, UK managing director of Tui, forecasted that AI will “completely transform” the travel industry, but many customers will still prefer to book their trips through human travel agents.

He said: “That is not going away anytime soon, in my view, because of that group of customers who value that. They use the technology a lot of those customers, but they still want to go in and look someone in the eye when they’re booking something. They value that trust element.”

Abta CEO Mark Tanzer said: “The challenge is to harness the potential which AI has to support our businesses, while continuing to celebrate and champion the value of the personal touch and expertise which comes with booking with a travel agent or tour operator.”

THE three jobs most at risk of being taken over by AI have been revealed.

British workers will be forced to compete with a surge of 100,000 virtual AI employees by the end of 2026, research by Agentic AI firm Gravitee has suggested.

1

Software engineers, HR managers and customer service managers appear to be most at risk.

Almost half of company executives surveyed by Gravitee said more than half the work of these roles could be done by an AI agent.

It means companies could cut these jobs to slash business costs.

The other jobs at risk include salesperson, marketing manager and communications manager.

The AI firm surveyed 250 executives at large UK companies and all said they were planning to use AI agents in some way.

Most of the companies said they would ’employ’ between 16 to 20 AI agents to carry out tasks previously carried out by human employees.

Meanwhile 65% of the firms said they would be making job cuts over the same period, suggesting much of the work done by their current employees could be performed more cheaply by AI bots.

Still, nearly a quarter of companies said they plan to increase their headcounts to support the AI expansion.

Rory Blundell, chief executive at Gravitee, said: “Agentic AI will offer a huge productivity boom to firms of all sizes, with AI agents completing rote, repetitive tasks while workers are freed up to focus on more interesting challenges.

“However it’s clear that, at least in the short term, there’s a risk of widespread job cuts until companies know how to balance the investment with the return, which may be shortsighted. Not least because AI Agents, without proper governance, can cause widespread chaos and disruption.”

A report by the International Monetary Fund (IMF) has warned that AI could hit three in five workers and even replace their jobs.

Companies and even the government have already begun to slash their workforce to make way for the technology.

The UK’s biggest accountancy firms – PwC, Deloitte, KPMG and EY – have reportedly slashed the number of junior roles available by up to 29 per cent, while the government is axing at least 10,000 roles in the civil service.

Amazon, Telstra, Salesforce and Fiverr have all announced plans for lay-offs in the last 12 months, with jobs going instead to AI agents. 

There were 315,550 available entry-level roles in November 2022 (when the popular AI tool ChatGPT was launched) across the UK, which dropped to 214,934 last month, according to the jobs site Adzuna.

The 40 roles most at risk

Similar research has previously been carried out by Microsoft, which revealed the 40 jobs most likely to be taken over by AI.

Researchers looked at 200,000 conversations between people in the US and Microsoft’s AI chatbot, called Copilot, as part of the study.

They found that the jobs most likely to be affected by AI are those in the communications industry.

Interpreters and translators are the jobs most at risk, followed by historians, and passenger attendants such as air hosts and hostesses or train ticket inspectors.

Sales representatives came fourth on the list, then writers and authors, and customer service representatives.

Other jobs on the list included radio DJs, concierge workers, teachers, public relations specialists and telephone operators.

Jobs that involve physically working with people, operating or monitoring machinery, or doing manual labour tend to be the least-impacted by AI.

This includes nursing assistants, sewage workers, painters and plasterers, roofers, cleaners, and massage therapists.

On Monday, the artificial intelligence giant OpenAI introduced a feature that would allow users to make purchases through ChatGPT, in partnership with Etsy and Shopify. Travel experts say this will pave the way for a big change.

The way we book holidays is set to change hugely as the travel industry braces for a “massively disruptive” tech update.

On Monday, the artificial intelligence giant OpenAI introduced a feature that would allow users to make purchases through ChatGPT, in partnership with Etsy and Shopify.

“Users can now buy directly from US Etsy sellers right in chat, with over a million Shopify merchants, like Glossier, SKIMS, Spanx and Vuori, coming soon. Today, Instant Checkout supports single-item purchases. Next, we’ll add multi-item carts and expand merchants and regions,” the organisation announced.

The move is set to disrupt the online commerce world, as soon the 700 million weekly users of ChatGPT will be able to shop on the AI interface, without visiting other websites. This will hand OpenAI a huge amount of power in the world of e-commerce, and make it a lot of money through fees from sellers.

READ MORE: Scotland’s ‘most haunted’ A-road famous for ‘phantom trucks’ and ghostsREAD MORE: Fears elderly will be stopped from flying as new Ryanair rule dubbed ‘ageist’

Following the announcement, experts in the future of travel have predicted a world in which ChatGPT – or another similar large-language model – will achieve market dominance and become a ‘super-app’, serving as a portal to every part of the internet.

It will be possible to plan holidays in great detail using the AI, as many tech-savvy travellers already do. The big difference in the future will be that it’d then be possible to book the entire holiday in a minimal number of clicks, without leaving the page, the experts predicted.

Speaking at the World Travel and Tourism Council (WTTC) summit in Rome on Tuesday, Guy Johnson, an anchor with Bloomberg, said: “That is a huge shift. That is a new model. That is massively disruptive.”

Gaurav Bhatnagar, co-founder of TBO.com, predicted that the changes would be positive from the customer’s perspective. “The quality of service will improve. You will need fewer people (as a travel company). From a customer’s perspective, it will be cheaper, easier to book, and the experience will be better,” he said.

The change would be far more significant for travel companies. “It is likely that when ChatGPT gets the customer to the point when they’re going to book, it will send the holiday plan out to their clients and say ‘who is going to bid for this?’ My guess is that ChatGPT wants to replace Amazon, Booking.com, to become the super app,”Mr Bhatnagar continued.

Paolo Benanti, president of the AI Commission for Information and a member of the UN Advisory Board on AI, was similarly boosterish. He said: “You would expect that an AI agent could arrange more accurately than a human can.”

The prospect of an AI revolution in travel has been brewing for a number of years. At last year’s WTTC summit, similar predictions were made about how large language models could be used to generate personalised travel itineraries for holidaymakers.

“Challenges that must be overcome by travel and tourism businesses include a shortage of AI-skilled workers, limited AI infrastructure, and the absence of a formal ‘AI strategy’ in many business plans. Travel and tourism businesses are urged to embrace AI as a strategic priority, investing heavily in talent to foster collaboration between humans and AI,” a statement from the WTTC read.

During the first months of the rollout of ChatGPT, the Mirror embarked on a mystery road trip arranged by the software and discovered some of the strong points and pitfalls of relying on a bot for advice and direction.

GOOGLE has always excelled when it comes to smartphone cameras – but the latest effort might have gone too far.

When I tried out the new Google Pixel 10 Pro XL’s star feature I gasped and said “Good God”.

6

6

I showed friends some of the results and they were stunned – if a little disturbed – by what the device’s mega zoom was capable of.

My colleague Sean Keach has already tried the smaller Pixel 10 Pro and has given his verdict on the identical camera features on there too.

But is it all worth it?

As The Sun’s main Android smartphone reviewer I’ve been putting the Pixel 10 Pro XL through its paces and here’s what I found…

Google Pixel 10 Pro XL: Look and feel

Google has stuck to a pretty consistent, solid look on the Pixel 10 Pro XL.

As ever with annual upgrades, you may even struggle to tell the difference between last year’s Pixel 9 Pro XL which I also reviewed.

I love the mat glass back and the shiny G logo, it just gives off such a confident premium vibe.

The camera bar is pretty much the same as last year too and I’m still undecided about the bulkiness of it, even though it’s quite tidy and neat.

But given the sorts of photos it takes, all that kit has to be packed in somewhere.

The display has had a noticeable upgrade this year, which is brighter and bursting with crisp detail too.

However, the Pixel 10 Pro XL hasn’t been taking a techy dose of Ozempic like Samsung and iPhone this year with thinner models – it weighs 232g, up a bit on last year.

I absolutely adore the new colours this year too, with Moonstone (the one I’m reviewing) Jade, Porcelain and Obsidian.

Google Pixel 10 Pro XL: Performance and features

Google continues to use its own Tensor chip in Pixel smartphones and this year’s Tensor G5 kit is said to be about 35 per cent faster.

It certainly operates smoothly and without any hiccups in my testing – but this chip isn’t as powerful as the Snapdragon chip found in most other top end smartphones around.

The handset runs on Android 16, the latest version of the operating system and it’s my absolute favourite around – which should be a given seeing as Google owns Android.

A lot of the features are centred on AI once again and most are photography based, which I’ll go into later.

But there are some other tools such as Voice Translate, which translates calls in real-time and sounds like each speaker’s own voice.

Having tried it with my partner in French, it was somewhat eerie to hear.

And it’s all done on device, so no sensitive call conversations are sent off to the cloud.

Google is also among the best for offering quick and fast Android upgrades for years to come, with a commitment to seven operating system versions and seven years of security upgrades.

Who offers free updates longest?

The longer you receive updates, the longer you can safely continue using your smartphone – with the latest features thrown in too for free.

Samsung
Samsung offers at least seven generations of OS updates and seven years of security updates on most of its smartphones, even its latest cheaper FE model.

OnePlus
At the launch of the OnePlus 13, OnePlus committed to at least four years of Android updates and five years of security updates.

Xiaomi
Xiaomi offers four years off Android updates and five years security updates.

Google
For the Pixel 10 series, Google said that devices would receive at least seven years of support.

Google Pixel 10 Pro XL: Battery

You can easily get a good day’s worth of use out of the Google Pixel 10 Pro XL with a good mix of browsing, messaging and Netflix as I tend to test it on.

But it’s not the best battery around – I got more out of the Galaxy S25 Ultra.

There is fast charging though, that’ll get you from zero to full in about an hour and a half with a compatible plug charger.

Google Pixel 10 Pro XL: Camera

The camera on the Google Pixel 10 Pro XL is where things get really juicy.

You have a treasure trove of powerful lenses paired up with AI power to enhance everything.

There’s a 50-megapixel main camera, 48-megapixel ultrawide and 48-megapixel 5x telephoto, as well as a 42-megapixel selfie snapper.

The photos I get on a Pixel are always super impressive with incredible detail and popping with vibrant colour.

It exudes beauty inside and out, and makes taking stunning photos effortless

This year’s shocker is the Pro Res Zoom which can get extremely detailed shots from a remarkable distance, seeing things my naked eye cannot.

It can go up to an eye-watering 100x.

Anything above 30x onward uses an AI model on the phone to recreate bits lost from digital zoom, so it does raise questions about how real the results are.

Whatever way you look at it, the images are highly convincing, as you’ll see from the snaps I took from the same position below with no zoom, a 10x zoom, a 50x zoom and a 100x zoom.

One thing to note: when using the zoom faces are automatically blurred.

6

6

6

6

Google Pixel 10 Pro XL: Price

One bit of good news about the price is it hasn’t increased from last year’s models.

So the Pixel 10 Pro XL starts at £1,199.

If that’s a bit steep for you, you might want to consider the Pixel 10 Pro instead which has much of the same in a smaller physical form from £999 instead.

There’s also a freebie worth £190 included with all the Pro models.

You get access to Gemini Pro for a whole year, which is Google’s more powerful and fast AI system, capable of extra skills such as Veo 3 Fast for text-to-video generation.

All prices in this article were correct at the time of writing, but may have since changed.

Always do your own research before making any purchase.