Insurance
Home›Insurance›Industry & Deals›OpenAI chief scientist urges voluntary slowdowns in AI…
OpenAI chief scientist urges voluntary slowdowns in AI development
Jakub Pachocki said he expects AI labs to coordinate voluntary delays until shared safety standards are established.
OpenAI chief scientist Jakub Pachocki warned that advances in artificial intelligence are accelerating so quickly that it is becoming increasingly difficult for humans to understand and control AI systems. Writing in a Sunday post, Pachocki said he expects AI models will soon be able to improve themselves without human intervention, a process he described as recursive self-improvement.
Pachocki said AI developers can respond by trying to steer models more closely toward alignment with human interests, or by slowing future development as needed. He added that he expects and hopes for voluntary slowdowns to become commonplace until shared safety “bars” are established, calling the current moment one that demands “extreme caution,” according to Insurance Journal.
The warning comes as AI labs continue releasing more capable systems, including OpenAI’s limited release of its GPT-6 Astra model, which the outlet said is tied to its advanced cybersecurity capabilities. The article also notes that Anthropic has not yet released its cutting-edge Mythos model publicly for similar reasons.
Insurance Journal said Pachocki’s concerns align with broader comments from other AI leaders, including Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman, who have each highlighted potential risks and the need for more communication around the technology’s impact. The piece further references claims that AI systems are becoming more capable at tasks such as research and cybersecurity, alongside growing investor interest in recursive self-improvement.