OpenAI Flags Possible Critical Cybersecurity Risk in Astra Model
OpenAI paused internal development and triggered safety protocols after acknowledging its upcoming Astra AI model may possess critical cybersecurity capabilities.
OpenAI revealed that it cannot rule out that its upcoming artificial intelligence model, Astra, possesses critical cybersecurity capabilities. The finding prompted the startup to pause some of its internal development activities and trigger safety protocols to manage potential risks associated with the software.[1][2]
Under the organization safety guidelines, an artificial intelligence model reaches the critical threshold if it can autonomously perform capabilities in the cybersecurity domain. Following the discovery, OpenAI has moved to tighten controls while evaluating the model.[1][2]
