OpenAI slows Astra development over cybersecurity concerns

OpenAI is slowing development of its Astra model today over cybersecurity concerns following an internal review, Engadget reports. The move highlights mounting industry caution about advanced AI systems that could be repurposed for offensive cyber tasks amid rising scrutiny from regulators and national security officials. It also raises questions about timelines for more autonomous AI agents and the guardrails needed before wider deployment.

According to Engadget, an internal assessment found the team could not rule out critical cyber capabilities, prompting a slowdown while risks are reviewed and safeguards strengthened. The outlet reported no revised timeline and said OpenAI has not detailed next steps, as the review focuses on potential misuse vectors and required controls; watch for an official OpenAI statement, updated milestones, and any regulatory outreach. [Developing. This article will be updated.]