OpenAI released SimpleQA, a new benchmark for generative AI. During the work, they uncovered a serious qualm about AI being supremely overconfident. Here's the scoop.

However, Yamin said that the chance of AI hallucinations and accuracy issues presents risks. Related stories Other companies are taking different approaches. OpenAI will continue to integrate election-related responses into ChatGPT while providing some ...

Sam Altman said OpenAI's closed approach offers "an easier way to hit the safety threshold" but he wants "to open source more stuff in the future."

There is a lot of uncertainty about artificial general intelligence, a still hypothetical form of AI that can reason as well — or better ... Miles Brundage, a former head of policy research and AGI readiness at OpenAI, told Hard Fork, a tech podcast ...

OpenAI CEO Sam Altman said that GPT-5, the company’s next probable large AI model, likely won’t come this year.

OpenAI is advancing the development of an inference chip with Broadcom but abandons plans to build a fab network.

In the meantime, the sources also said OpenAI is incorporating AMD chips into its Microsoft Azure setup. AMD introduced its MI300 chips last year, which was a big part of the news this summer that its data center business has doubled in a single year as it chases market leader Nvidia.

Kalinowski posted the announcement on X and people were immediately excited by the news, with speculations over whether this would lead to an OpenAI robot or a more consumer-friendly device such as a pair of smart glasses.

In an event hosted by and streamed to Bilibili, researcher Lilian Weng Li called AI a ‘double-edged sword’ that must be trained properly.

Meta makes its model available to the U.S. military with a pitch that's largely about defending its "open source AI" framing.