OpenAI released the GPT-6 Astra model on September 3. The company assigned the model a Critical rating for cybersecurity capabilities under its Preparedness Framework. According to the documentation, the model can identify unknown vulnerabilities and develop exploitation methods with minimal human intervention.
OpenAI reported that it slowed down specific development stages for Astra to enhance safety measures. Following this release, leaders of major AI companies, including Sam Altman, Dario Amodei, and Elon Musk, have begun discussing the need to slow down the pace of AI development. They cite concerns that model capabilities are outpacing current safety systems.
The industry shift toward slower development has impacted the market. OpenAI has reportedly delayed its IPO, and discussions regarding potential restrictions on AI development have negatively affected the stock prices of companies associated with the artificial intelligence sector.
