OpenAI scraps rollout of latest model amid safety concerns
The decision follows calls from industry leaders for a slower pace of AI development and tougher safety safeguards
OpenAI has scrapped the planned release of its newest AI model after internal testing found it failed to meet the company’s safety and alignment standards.
Listen to this article
GPT-6.1 Astra, a next-generation model expected to launch in October, had been due to power more complex tasks across ChatGPT and Codex with less human input.
The decision follows calls from OpenAI chief executive Sam Altman, Anthropic CEO Dario Amodei and other industry leaders for a slower pace of AI development and tougher safety safeguards.
OpenAI has previously warned that Astra could, at times, evade human oversight.
The company and rivals including Anthropic have also faced scrutiny over experimental systems that breached safeguards, including reports of an OpenAI model accessing Australia’s health system database.
Read more: Trump to hold AI summit despite branding fears a 'hoax'
Read more: Pope takes aim at Trump over AI as pontiff and president reignite feud
The Journal reported that GPT-6.1 Astra also showed higher levels of deception than its predecessor in internal testing, including instances in which it did not always accurately disclose what actions it had taken.
"While (GPT-6.1 Astra) improved on axes such as model laziness, it didn't quite meet the bar in terms of staying within scope and authorization, and how it communicates back to the user about the type of work it's done," said Saachi Jain, head of safety systems at OpenAI.
"Of course we want to make sure our model development is safe no matter whether that's in the company, or when we ship it to users. But when we ship it to users, we have an extremely high bar in terms of safety and alignment," Jain said.
The decision comes ahead of OpenAI's developer conference in San Francisco, where the company has previously unveiled products aimed at software developers.