Models & Labs · United States

OpenAI’s math solutions aren’t meeting the field’s standards yet

Credited to TechCrunch · techcrunch.com · 16h ago

When OpenAI released hundreds of claimed solutions to some of the world’s hardest math problems this week, the frontier lab said that it had consulted an advisory group of elite mathematicians to avoid the controversy that came with the last time one of its models solved a long-standing problem in the field.

But OpenAI fell short of those standards, particularly where the mathematicians emphasized the need for human understanding of a mathematical result. That’s especially concerning after a new paper highlighted gaps between the natural language and formally expressed solution to a million-dollar problem ostensibly solved by OpenAI’s models.

The Advisory Group on Mathematics and Artificial Intelligence (AGMAI), hosted by Princeton University’s Institute for Advanced Studies, is made up of nine prominent researchers at institutions around the world.

The organization released guidelines for frontier labs solving math problems at the end of September. In a statement on the latest set of proofs, the AGMAI said that “it is ultimately up to the mathematical community to assess the extent to which our recommendations were followed successfully.”

However, the organization’s first request was “to stop testing advanced mathematical problems on proprietary models.” OpenAI’s release explicitly says that it is evaluating its proprietary models using open research problems in mathematics.

The advisory group did not respond when asked by TechCrunch for a more thorough evaluation of OpenAI’s latest proof release. The lab clearly followed some of its principles, including releasing results as soon as possible and including information about how the models reached their conclusions. But not for all of them: Just 10 of the 719 manuscripts included releases of the model’s chain of thought.

For papers that people don’t understand, the mathematicians suggested the proofs should be formalized — but just 42% of the proofs released by OpenAI had not undergone this process.

Ultimately, it’s still not clear that OpenAI is taking “responsibility for ensuring that human understanding will follow” when releasing its proofs, in accordance to the AGMAI principles. AGMAI suggested that OpenAI should help fund the work of human mathematicians who will be required to make the lab’s solutions meaningful in any real way.

“Problems are being solved autonomously by AI prompters who have no interest in the broader field itself once their initial target is ‘solved’, and do not understand the AI output well enough to answer questions on the result, give talks, or otherwise interact with the rest of the field,” Terence Tao, a prominent mathematician who has criticized OpenAI’s approach, wrote on social media after the release.

That problem is exemplified by a paper released this week by mathematicians at the University of Cambridge and King’s College in London that questions the way frontier labs are approaching these challenges.

When AI models solve mathematical problems, they first create a “natural language” explanation, then try to express that result in Lean, a programming language that in theory confirms the accuracy of the proof by compiling it as code.

However, there may be problems with the way the models translate their natural language proofs into code; this paper documents at least two discrepancies between the natural language proof and the Lean code behind the solution OpenAI has offered to a problem derived from the Navier-Stokes equations that describe the complex behavior of fluids.

These discrepancies don’t necessarily disprove either solution, but they do raise questions on whether we can simply rely on models to formalize their own solutions without human involvement. That’s one reason that AGMAI asked OpenAI to “include machine-readable metadata correlating the natural language and formal artifacts,” something that the frontier lab did not do with these releases.

“Because of the phenomenon of mistranslations — as highlighted in this paper — the NL proof by OpenAI and other autoformalised Lean proofs should not prima facie be trusted without the same peer review process and scrutiny that other proofs are subjected to,” the authors of the “lost in translation” paper conclude.

Mathematicians stress that when new results are discovered by humans, they take responsibility for them and engage with the broader community through papers, talks, and seminars. That process increases understanding of the solutions, finds strategies that can be used to solve other problems, and allows the new knowledge to be applied in practical fields.

When a model is prompted to solve a hard problem and spits out a solution, “there is not human understanding of them at the point of release, and now the work begins,” Harvard University mathematics professor Melanie Wood told TechCrunch.

Models & LabsUnited States

Why JPMorgan Is the Most AI-Advanced Bank

Credited to Bloomberg$ · bloomberg.com2h ago

Useful

Artificial Intelligence deployment across banking is moving at its fastest pace yet, with JPMorgan yet again leading Evident Insight’s ranking of the world’s most AI-advanced banks. Evident CEO and co-founder Alexandra Mousavizadeh explains why early movers are pulling ahead and…

View the article · $Bloomberg

Front Method

Models & Labs · United States

Iran's supreme leader calls for greater use of AI to strengthen security

Credited to The National · thenationalnews.com · 2h ago

Photograph · The National

Iran's supreme leader Ayatollah Mojtaba Khamenei has called for greater investment in artificial intelligence and advanced technology to strengthen the country's security forces.

In a message marking National Police Week, he said Iran needed to bolster its police force through improvements in personnel, equipment and technology, particularly the use of AI in data analysis.

"A comprehensive strengthening of the forces responsible for safeguarding security in all respects, including human resources, hardware and software – particularly through the advancement of technology, such as the purposeful use of artificial intelligence in data analysis – plays an important role in improving the ability of the police force to carry out their missions and facilitates their work," he said.

Mr Khamenei also praised the country's security forces for maintaining order despite repeated US and Israeli attacks. He accused the US and Israel of deliberately targeting police commanders and personnel, but said the attacks had failed to undermine the ability of security forces to carry out their missions.

The National revealed late last year that Iran planned to use AI in its armed forces and security apparatus – nearly three months before the US and Israel launched strikes on Iran in late February, sparking the regional war.

Despite restrictions on its access to advanced AI chips owing to sanctions, Iran's Islamic Revolutionary Guard Corps said last year during military exercises that it was using AI in domestically developed drones, including in missiles capable of autonomously locking on to moving targets.

The announcement highlighted Tehran's determination to keep pace with the global AI race. Under an official strategy, Iran aims to rank among the world's top 10 AI powers by 2032, with AI contributing 12 per cent of its GDP.

The push to integrate AI into Iran's defence systems accelerated after its first direct military confrontation with Israel in 2024. In April, Brig Gen Kioumars Heydari, commander of the army’s ground forces at the time and now deputy chief of staff, described AI as "one of our priorities to maximise our defensive power”.

The US has also used AI in its war against Iran. In March, Admiral Brad Cooper, commander of US Central Command, said AI systems “allow us to analyse large volumes of data within seconds, enabling our leaders to make the right decisions faster than the enemy amid the noise”.

He said advanced AI tools had reduced processes that would typically take hours, and sometimes days, to mere seconds. “However, the final decision is always made by a human: what to strike, what not to strike, and when to strike,” he added.

But the US bombing of an elementary school in Minab, southern Iran, on the first day of the war – an attack that killed more than 150 people, many of them schoolgirls – raised questions about AI's role in military operations.

A September report by Bloomberg said an overreliance on AI tools, alongside outdated intelligence and imagery, contributed to the strike. The full findings of a Pentagon investigation into the attack have not yet been made public.