openai

OpenAI’s announcement is the latest in a series of claimed scientific breakthroughs credited to AI.

OpenAI has announced that its AI models have made several scientific discoveries in mathematics, computer science and physics.

The AI giant said on Monday that the three new discoveries concerned the fields of harmonic analysis, complexity theory and quantum field theory.

Among the discoveries, AI models proved the local smoothing conjecture, one of the central propositions of harmonic analysis, in every spatial dimension for the first time, according to OpenAI.

Researchers in harmonic analysis, a branch of mathematics that studies how complex signals and functions can be broken down into elementary waves, had previously only proved the theory in three dimensions.

‘Historic development’

Alex Iosevich, a professor of mathematics at the University of Rochester who specialises in harmonic analysis, said OpenAI appeared to have “cracked” the conjecture in a “historic development”.

“The conjecture was formulated by my PhD adviser, Chris Sogge, so it is near and dear to my heart, and the news was a shock at first,” Iosevich told Al Jazeera.

“Now I am filled with joy and optimism, because the scientific enterprise as a whole is going to make giant leaps as a result of the AI revolution, and we have the chance to be a part of this historic development.”

Iosevich said he expected the field of theoretical mathematics to “change drastically” due to AI.

“I feel a twinge of sadness as a result,” Iosevich said.

“Many theoretical mathematicians I know feel the same way, but I firmly believe that these feelings will soon give way to the realisation that there is so much we can now accomplish.”

“Problems in pure mathematics did not arise in a vacuum,” he added. “Many of them came out of concrete physical considerations, and these connections can now be explored rapidly and decisively.”

According to OpenAI’s announcement, its models also disproved the Exponential Time Hypothesis, which relates to the time it takes to solve complex computational problems, and provided “mathematically rigorous [proof of the] existence” of quantum chromodynamics, which studies the interaction of quarks and gluons.

The ChatGPT creator said it had reached out to scientists outside the company to turn its “raw research data into scientific understanding”.

The San Francisco-based company said it decided on several different methods for releasing the findings based on consultations with three expert groups chosen to act as “shepherds”.

While the groups collaborating on the mathematics and computer science problems opted to make the AI-produced data public immediately, the physics group decided to defer the release in order to write up their own interpretation of the findings, OpenAI said.

“All of these approaches involve tradeoffs,” OpenAI said.

“On the one hand, the collaborations empower experts to foster unpacking the results and contextualizing them in the broader academic community,” the $852bn start-up said.

“On the other hand, choosing initial shepherds necessarily shapes who gets involved first and creates asymmetry of information. We welcome feedback on how to broaden participation and make this process more useful.”

OpenAI’s announcement is the latest in a series of claimed scientific breakthroughs credited to AI.

Last month, Claude creator Anthropic announced the discovery of a new enzyme said to resemble CRISPR, the groundbreaking gene editing mechanism used in treating conditions such as cancer and sickle cell disease.

Anthropic’s announcement drew mixed reactions among scientists.

While some researchers hailed the finding as an exciting development, others suggested that the AI company had exaggerated its significance.

Americans can rest easy – for the moment, at least. The fine folks at the Federal Bureau of Investigation have thwarted a major national security threat.

On October 9, the FBI arrested 35-year-old engineer Abdulrahman Alfozan, a Saudi citizen employed at the California-based artificial intelligence company OpenAI, the maker of the chatbot ChatGPT, and a firm that describes its mission as ensuring that “artificial general intelligence benefits all of humanity”.

A press release from the United States Department of Justice said that Alfozan was accused of installing a “surreptitious audio recording device” in an executive conference room at the OpenAI offices in San Francisco.

The statement quoted Director of National Intelligence Jay Clayton’s assessment that “this serious situation underscores why winning the global super intelligence race is a national security prerogative”. (For those who have not yet got the memo, US President Donald Trump recently decreed that the term “artificial intelligence” be changed to “super intelligence” or “SI.”)

FBI director Kash Patel concurred that “protecting American SI innovation is essential to safeguarding our nation’s economic strength and national security”, and assured everyone that the FBI would continue doing its part “to ensure our country maintains its competitive edge”.

So much for the separation of corporation and state in the world’s self-appointed greatest democracy.

The covert device was reported to have been concealed in toilet paper and attached to the back of a television monitor. The Gizmodo website has noted with amusement that, despite being cast as a “sophisticated” spying apparatus, the device in question can in fact be easily ordered online “for the low low price of $13.99”.

If convicted, Alfozan faces up to five years in jail and a fine of $250,000.

Now, some of you may still be asking what the big deal is – and why the US government is acting like it’s 1950 and the Soviets are stealing our nuclear secrets.

The spectacle becomes all the more curious when we consider that the company up in arms over the alleged attempted theft of its technology owes much of its existence – and massive valuation – to stealing stuff from people.

The day before Alfozan’s arrest, USA Today and several other newspapers sued OpenAI in a Manhattan federal court for copyright infringement and the unauthorised use of hundreds of thousands of news articles and other materials to develop ChatGPT.

And back in July, various US media outlets including the New York Times and the New York Daily News requested that a federal judge impose sanctions on OpenAI for “hiding and destroying evidence showing how ChatGPT was trained on stolen journalism”, as attorney Steven Lieberman put it.

The moral of the story, perhaps, is that when OpenAI steals things it’s to “benefit all of humanity” – but when a 35-year-old Saudi engineer does it, it’s a national security crisis.

In case we were not already bored to tears hearing about national security, US Attorney General Todd Blanche took to X on Tuesday to preach that the “theft of intellectual property involving cutting-edge super intelligence technology poses a serious threat to our national security”, necessitating continued “swift action … to protect American innovation and safeguard our nation”.

Of course, many Americans might wonder why “innovation” gets such special treatment when they themselves are so often on the receiving end of government abuse. The frequency with which US police officers murder unarmed Black people, for instance, would seem to indicate that there are far more literal threats to the nation’s security than a $14 device hidden in toilet paper.

But human security is of no interest to a corporatocracy increasingly dominated by AI, an industry that thrives on putting people out of jobs and that is deeply entwined in a number of sordid developments at home and abroad. Among them is mass surveillance, which combines AI-assisted violations of civil liberties and constitutional protections with good old police state “innovation”.

To that end, a bunch of US cops have taken advantage of new AI-powered technology, such as automated licence plate readers, to stalk their exes. Then there is the matter of the FBI and other government agencies buying up data en masse from tech companies in a bid to keep tabs, Big Brother-style, on everyone’s personal information at all times.

It makes you think twice about the sudden concern for “spying” in the Alfozan case.

As for threats to other people’s national security, OpenAI and like-minded tech companies that do business with the Pentagon have been complicit in recent US aggression – from Venezuela to Iran – as the US pursues its strategy for “Accelerating America’s Military AI Dominance” (concocted before Trump mandated the renaming of AI as SI).

Israel, too, has exploited OpenAI technology, among other US-made goodies, to conduct its contemporary apocalyptic exercises in the Middle East. Last year, the Associated Press news agency (AP) revealed that US tech giants had “quietly empowered Israel to track and kill many more alleged militants more quickly in Gaza and Lebanon through a sharp spike in artificial intelligence and computing services”.

The AP report went on to note that “the number of civilians killed has also soared, fueling fears that these tools are contributing to the deaths of innocent people”.

But while it might be news to the AP, that is unfortunately how a genocide works.

And as the US goes to war for AI dominance while the tech giants bank on a post-human world, one can’t help but feel that the toilet paper was a nice touch, indeed.

The views expressed in this article are the author’s own and do not necessarily reflect Al Jazeera’s editorial stance.

United States federal agents have arrested a Saudi national and charged him with planting a hidden audio recording device in a conference room at the California offices of OpenAI, where he was working as an engineer.

Abdulrahman Alfozan, 35, had been working at the AI company since ⁠July 2025, the US Department of Justice said in a statement on Tuesday.

OpenAI discovered the device plugged into a USB port behind a television in an executive conference room at one of its San Francisco buildings on September 30, according to a criminal complaint filed by the Federal Bureau of Investigation (FBI).

Following the discovery, the company placed a camera in the room, which captured Alfozan entering last week and checking behind the television where the listening device had been installed, the FBI alleges.

The case highlights growing security concerns around sensitive AI research as companies and governments compete for the best talent and technology with the potential to alter regional economic and strategic dynamics.

Who is Abdulrahman Alfozan and what is he charged with?

Alfozan has been a legal permanent resident of the US since September 2020.

According to a LinkedIn profile in his name, he studied at ⁠the Massachusetts Institute of Technology (MIT) and the University of California, Berkeley. He previously worked at Meta and Microsoft, it said.

Before he joined OpenAI, he was a venture partner at Raed Ventures, headquartered in Riyadh in Saudi Arabia, which invested “in seed and early-stage startups in Saudi Arabia and MENA”.

Al Jazeera could not independently verify the details mentioned in this LinkedIn profile.

Alfozan made his first court appearance on Tuesday on a charge of “using or endeavouring to use an electronic, mechanical, or other device to intercept oral communications”.

It was not immediately clear if he has legal representation. He is expected to return to court on October 19.

If convicted, Alfozan faces up to five years in prison, a $250,000 ⁠fine, and three years of supervised release.

What happened at OpenAI?

The FBI says a thumb-sized device with a microphone, and which ‌could connect to a camera, had been concealed in toilet paper and placed behind a television set in the conference room. It was installed around September 28, the FBI said. Its stored data was encrypted and inaccessible. Investigators later found a similar device in Alfozan’s personal bag.

In a post on X, OpenAI said it had evidence that the employee intended to place other listening devices throughout sensitive areas of its offices to secretly record meetings.

“At this time, we have no indications of this individual inappropriately accessing or exfiltrating model weights, source code, or customer data. The employee was ⁠terminated,” the company said.

The FBI complaint also alleges that Alfozan used an AI productivity tool to make multiple inquiries about the company’s building addresses, meeting rooms, meeting-room capacities and calendars, despite not having the authority to access these spaces or participate in meetings there.

“This alleged attempt by a Saudi national to steal sensitive communications from OpenAI was taken extremely seriously. Theft of intellectual property involving cutting-edge super intelligence technology poses a serious threat to our national security,” Attorney General Todd Blanche said in a statement on X.

Hasn’t OpenAI been accused of helping to spy on civilians?

In an X post by OpenAI explaining how its own systems were used to trap Alfozan, the company said: “Our security program is designed to protect against a wide range of threat actors, from insider risks to cybercriminals to nation states, and this case is an example of how AI can help address these sophisticated security threats.”

But AI technology is used not only to catch spying, but conversely, to engage in it as well.

In February 2025, an investigation by The Associated Press found that US tech giants had enabled Israel to track and kill people in Gaza and Lebanon through AI and computing services. The AP reported that the Israeli military had bought advanced AI tools from tech companies including OpenAI, the maker of ChatGPT, through Microsoft’s Azure cloud platform, to assist with surveillance.

OpenAI denies having a relationship with the Israeli military.
But OpenAI also faced criticism when it signed a deal with the US Department of Defense in February this year to supply its technology to the US military for use in classified settings, including intelligence gathering. OpenAI’s competitor, Anthropic, refused to sign the same deal unless the Pentagon barred mass surveillance of Americans.

There have been several incidents in which OpenAI models have gone “rogue” by independently breaking out of a testing environment and then hacking into other organisations and even government sites, leading to calls for stronger regulations against unchecked AI surveillance.

OpenAI has admitted its actions were “not good enough” after one of its rogue agents breached Australian government websites in June and apologised again.

The company’s chief strategy officer Jason Kwon faced a parliamentary hearing into AI on Tuesday in Sydney, saying the breach “should not have happened” and it “should have handled our response better”. It took weeks before Australia was notified via an email to a generic inbox.

“We are sorry and we know we have work to do to rebuild trust with the Australian people,” Kwon said.

Anthropic also appeared at the hearing and said it had carried out a “lengthy, deep investigation” in recent months and “not found any cases” of Australian breaches.

An OpenAI agent went “rogue” and “infiltrated” a private statistics portal containing “non-sensitive” data from Australia’s universal healthcare scheme Medicare in June in what cyber-security experts said was the first hack of its kind.

Asked why OpenAI had not directly contacted government ministers immediately after it found out about the breaches, Kwon acknowledged it was a mistake.

“On retrospect, we should have done what you’re suggesting,” Kwon said, in response to a question by the committee on why it had not dialled the mobile numbers of government ministers.

He said the company has changed how it handles such incidents.

“Even if we don’t fully understand the situation, we are just going to notify and start working through the sitation collaboratively with the impacted party.”

OpenAI has also added “more precautions” to its training environments since the incidents, Kwon told the 12-member committee which is made up of Labor, Liberal and independent MPs who are looking at AI and it’s impact on Australia.

The company was also establishing a local taskforce in Australia to investigate “how to better manage the risks associated with increasingly capable AI”.

OpenAI has fired three researchers for allegedly mishandling information, including work that involved an external organisation that analyses artificial intelligence (AI) models.

“Our investigation confirmed that these individuals mishandled sensitive information outside established company procedures, violating our policies and breaking the trust essential to our work,” a spokesperson told the BBC.

The ChatGPT-maker did not name the sacked workers, but at least two of them were involved in safety research at the firm.

The firings come as the debate over AI safety and the risks the technology may pose to humanity has intensified in recent weeks.

“We have parted ways with three individuals for violating our policies on accessing and handling sensitive company information,” OpenAI’s spokesperson said.

The BBC understands the former employees were not let go for raising safety concerns but for allegedly mishandling sensitive information.

Concerns over AI safety have made headlines after some researchers and industry executives called for more guardrails around the technology.

OpenAI has come under under intense scrutiny after its models went rogue and hacked several platforms, including Australian government websites.

In an earlier incident an OpenAI system accessed the internet and breached the open-source developer platform Hugging Face.

That incident led OpenAI to conduct a broad review of the activities of its AI models – called agents, which are designed to execute tasks autonomously based on simple instructions.

The AI lab said this week that it has notified more than 100 organisations about incidents involving unauthoritised activity linked to its AI systems.

Being notified “does not mean that any private information was accessed” or that a system was compromised, OpenAI said.

DEVELOPING STORYDEVELOPING STORY,

OpenAI has cancelled the release of its latest AI model over safety concerns, according to media reports.

The move to scrap the release of GPT-6.1 Astra, reported by the Wall Street Journal and CNN on Monday, comes amid heightened concerns about AI’s potential to do catastrophic harm following a slew of incidents involving AI agents going rogue.

More to follow…