OpenAI's chief scientist warned that preparations for AI's rapid progress are inadequate. Following the release of the latest model and reports of attacks by autonomous agents, stronger measures going forward have been suggested.
OpenAI chief scientist Jakub Pachocki (ヤクブ・パホツキ) called for an extremely cautious response to AI's rapid rise and said in a blog post that more intervention may be needed to ensure control over the future.
Shortly beforehand, the company released its latest generative AI model, GPT-6 Astra, describing it as its most powerful product to date. Autonomous AI agents associated with OpenAI and Anthropic have reportedly carried out cyberattacks against other companies in the real world.
To ensure AI safety, Jakub Pachocki (ヤクブ・パホツキ) proposed establishing legal and international minimum standards and third-party oversight bodies, and urged companies to proceed with development only if they meet safety standards. He also identified how to balance AI research with human involvement as a challenge.