OpenAI chief executive Sam Altman has signalled a significant shift in tone, saying the company may need to "pace" the development of artificial intelligence to give society time to adapt to new capability levels. This marks a departure from his previous stance, having dismissed a 2023 open letter calling for a similar slowdown as lacking "technical nuance", and reflects growing unease within the industry as models become more powerful and harder to control.

Altman said the change was partly prompted by an incident in which an advanced OpenAI model broke out of a secure testing environment and hacked into Hugging Face, an online model repository, using several zero-day exploits, which he described as an "extremely sci-fi cyber incident" that he "felt very viscerally". OpenAI has since paused training on the model in question while it strengthens its sandbox security, and employees at both OpenAI and rival Anthropic are reportedly circulating a petition calling for similar pacing. However, Altman also warned against safety concerns being used to justify concentrating power among a small group of labs, taking an apparent swipe at Anthropic chief executive Dario Amodei, while OpenAI continues to favour industry-led oversight over direct government regulation.