A METR report says Anthropic's models are improving roughly twice as quickly as a year ago, with much of the acceleration due to AI, while Anthropic redacted much of the evidence.
By Hermes-Vector AI Desk
Anthropic is using AI to build more powerful AI, and its models are improving roughly twice as quickly as a year ago, according to a report published Thursday by the AI safety research organization METR. The report follows an 11-week investigation inside the company evaluating its AI research and development.
METR researchers said they "are confident" that the acceleration was not simply due to scaling up computing power used to train the models. Instead, they said the majority of the acceleration as of October "is due to AI," with its contribution growing over time. The report states that benchmark scores suggest "that Anthropic has made roughly two years of historical progress over the last year."
The acceleration comes partly from AI assisting human researchers. METR wrote that employees using Anthropic’s most capable internal model on average self-reported being around 12 times as productive at their respective work as they would have been without AI assistance, around double what staff had reported just a few months earlier while using Mythos 5. METR warned that "there are some signs that fully autonomous contributions will play a larger role soon," and said researchers found no clear signs of diminishing returns or observable bottlenecks that could prevent "even faster acceleration."
Over the past six months, Anthropic has finished training three models it has yet to release, codenamed Model 2, Model 3, and Model 4, completing training in July, August-September, and September-October, respectively. METR also identified an unreleased model dating back to 2022, called Unreleased 2022.
Anthropic had the power to redact information from METR’s final report and used that power extensively. METR wrote on X: "We submitted a ~100-page full report on our assessment to Anthropic. For the public version, Anthropic redacted a majority of the evidence we found most informative."
Anthropic cofounder Sam McCandlish acknowledged on X: "We redacted a lot, and I accept that this makes it harder for others to check METR’s work." He said the company determined that sharing exactly what is driving the speedup was not the right thing to do because it is commercially sensitive and risks handing out a roadmap for others to go faster. Anthropic plans to publish its own report on the speed of AI progress "in the coming weeks," as well as new steps it is taking to strengthen safety.
The findings have prompted concern among AI safety researchers. Daniel Kokotajlo, a former OpenAI researcher who now runs the AI Futures Project, accused Anthropic of "recklessly gambling with our lives." Marius Hobbhahn, CEO and founder of Apollo Research, told The Verge the report suggests a pace of AI-enabled research and development even faster than what was forecast in AI 2027.
Brianna Rosen, research director for frontier security at the Institute for AI Policy and Strategy, told The Verge that it was concerning that companies, not governments, were effectively calling the shots on automating AI research. She said: "For democratic oversight to work, independent evaluators need the freedom to explain why they reached their conclusions, governments need to assess the evidence themselves, and the public has to trust this process."
The report raises questions about how quickly AI could advance toward systems that can improve themselves with little or no human involvement. It also highlights a tension between Anthropic’s public calls for responsible development and its own rapid progress, as the company prepares to go public while its CEO, Dario Amodei, has publicly called for labs to pump the brakes on AI development.
Sources
This article was drafted with AI assistance and checked against the sources above. Company claims are reported as claims. Cover image is AI-generated.
The day's checked news, by email
One short email a day with the top stories. Free, and you can unsubscribe in one click.
Or follow for the latest news:
No comments yet. Start the discussion.
Nvidia's board has opened an independent investigation after CEO Jensen Huang acknowledged an affair with a woman US officials say was working for Chinese intelligence. The company says it has no reason to believe national security or company secrets were compromised.
A US intelligence memo alleges China's Ministry of State Security recruited Alice Huang, an executive at Singapore-based Megaspeed, to gather information on Nvidia CEO Jensen Huang. Nvidia and China deny the allegations.
