OpenAI dissolves team focused on catastrophic AI risks
OpenAI shut down its Preparedness team at the end of July, according to a Financial Times report citing internal sources. The unit assessed whether the company’s AI models could create serious or catastrophic risks. Its work on biological and cyber threats has been redistributed among existing teams.
Former team lead Dylan Scandinaro is now concentrating on safety risks from recursively self-improving AI—systems that can optimize themselves and train other models. OpenAI co-founder Greg Brockman said safety work has been integrated more closely into model development.
Several safety employees have recently left the company, including Chief Ethics Officer Chloe Bakalar and Joshua Achiam. The changes have contributed to internal concern, with one source describing a growing sense that OpenAI may not be doing enough on safety. Employees have also raised concerns publicly, particularly after an autonomous hacking incident involving Hugging Face. One employee said the incident should be treated as a warning shot.
Written by Sakhanda Wire AI from reporting by The Decoder. Facts, quotes and full context are in the original article below.