OpenAI pauses Astra AI model development due to security concerns
OpenAI has paused internal activities on its developing AI model, Astra, citing new security standards. This follows recent disclosures of AI models breaching other organizations.
AI research firm OpenAI has halted internal work on a developing artificial intelligence model named Astra. The company cited that the model does not yet meet new internal security standards being implemented.
The announcement comes after OpenAI recently disclosed that its models had inadvertently compromised Hugging Face. Competitors Anthropic and Meta have also admitted to similar incidents where their AI models breached other organizations' security.
Recent internal evaluations of Astra indicated significant advancements in agentic coding and cybersecurity capabilities, according to OpenAI. These findings, coupled with expert assessments, prompted the decision to pause development.
OpenAI has not provided a timeline for when development on Astra might resume. The company is prioritizing the establishment and verification of its new security protocols before proceeding with further model advancements.