OpenAI Puts New AI Model Under Tighter Controls As Its Cyber Capabilities Soar
OpenAI scaled up security controls and paused internal Astra work after preliminary evaluations suggested the model could reach a critical cybersecurity threshold.
- On Friday, OpenAI paused internal development of its upcoming AI model, Astra, after preliminary evaluations could not rule out "critical" cybersecurity capabilities, triggering enhanced safety protocols.
- OpenAI attributed the issue to a "misconfiguration" within Irregular's testing ground, an Israeli startup whose platform "allowed models to access the public internet" during evaluations.
- In recent weeks, OpenAI, Anthropic, and Meta Platforms disclosed their AI models broke into systems during cybersecurity testing, with incidents linked to the "same evaluation-environment issue" first identified by Anthropic.
- Sundeep Bhimireddy, head of AI at Von, criticized the current testing approach, stating, "When they are testing these models, they don't want to grade their own homework," and advocating for independent third-party vendors.
- Democratic Rep. Ted Lieu of California told CNBC this week, "We need to get this bill across the finish line this year," citing "unauthorized hacks of other companies" as pressure mounts for federal oversight.
30 Articles
30 Articles
OpenAI Pauses Some Astra Work Over Critical Cybersecurity Concerns
OpenAI is pausing some internal work involving its upcoming Astra AI model after preliminary evaluations raised concerns that it could approach the company’s highest cybersecurity capability threshold. In its official Astra cybersecurity disclosure, OpenAI said recent internal evaluations showed significant advances in agentic coding and cybersecurity. The results were strong enough that the company said it “cannot rule out critical cyber capabi…
OpenAI Puts New AI Model Under Tighter Controls As Its Cyber Capabilities Soar
OpenAI said Friday that its unreleased model, Astra, has performed strongly enough in cybersecurity evaluations that the company cannot yet rule out the possibility that it has reached "Critical" capability.
The company has implemented stricter security measures and is working with government agencies and individual organizations in the field of AI security for additional testing.
OpenAI Astra Model Development Paused Over Cyber Risks | 📲 LatestLY
OpenAI has paused internal development on its upcoming AI model, Astra, after evaluations could not rule out critical cybersecurity capabilities. According to a Business Standard report, the model showed advanced autonomous potential to exploit software vulnerabilities, prompting stricter sandboxed security controls. 📲 OpenAI Astra Model Development Paused Over Cyber Risks.
OpenAI pauses some work on new Astra model on cyber concerns
OpenAI is pausing some internal work around one of its upcoming artificial intelligence models to implement stricter safeguards after the system was found to be significantly more adept at cybersecurity tasks.
Coverage Details
Bias Distribution
- 55% of the sources lean Left
Factuality
To view factuality data please Upgrade to Premium
