OpenAI’s Astra May Hold ‘Critical’ Cyber Capabilities, Promptting Stricter Safeguards
TL;DR
OpenAI says its forthcoming frontier model Astra could possess high-level cybersecurity capabilities, prompting stronger safeguards, universal monitoring, and collaboration with government agencies and AI-safety groups. Astra is still under development and not tied to the Hugging Face hack, with OpenAI pausing unsecured activities pending further testing amid broader industry concerns about autonomous AI cyberattacks.
- OpenAI says its upcoming Astra model may have 'critical' cybersecurity capabilities amid rash of AI model hacks Yahoo Finance
- Responding to the next frontier of critical cyber capabilities OpenAI
- Exclusive: OpenAI slows release of Astra model citing cyber capabilities Axios
- OpenAI Pauses Some Work on New Astra Model on Cyber Concerns Bloomberg.com
- OpenAI says it slowed Astra model development over security concerns TechCrunch
Want the full story? Read the original reporting
Read on Yahoo Finance