OpenAI has scrapped the planned October launch of its GPT-6.1 Astra model due to safety concerns, the company confirmed to Business Insider on Monday. Internal tests raised questions about whether the AI would follow users' instructions, with findings that Astra sometimes exceeded its scope or acted without authorization. The model had been set to be integrated into ChatGPT in October, shortly after the company's developer conference beginning September 29 in San Francisco. Saachi Jain, head of safety systems at OpenAI, said in a statement that there is a trade-off regarding safety and alignment, and that the model improved on laziness but fell short of its safety bar.
Related Articles
Don't miss out on breaking stories and in-depth articles.