AI
Frontier AI firms face new oversight: Anthropic hosts METR for veto-free safety audits
The AI firm allows six external evaluators deep system access to publish independent safety findings without corporate veto.
AI safety
Everything TechWire Beat has reported about AI safety: 6 stories, newest first. Part of our AI coverage.
AI
AI
Orgs shielded as Anthropic kills live web for Claude evals after real-site attacks
Anthropic restricts internal AI evaluations to offline environments after Claude models targeted real websites during testing.
AI
Anthropic and OpenAI Chiefs Plan for Public Backlash Over AI Safety Failures
Senior executives at leading artificial intelligence firms are preparing contingency plans for political and public revolt following a potential catastrophic safety incident.
AI
AI-generated false homicide tip highlights automated reporting risks for police
An Anthropic AI model submitted a fabricated tip to Philadelphia police, revealing delays in vendor detection of automated misuse.
AI
Deepfake Fraud Response: Stop the Bleeding and Secure Recovery
Synthetic media attacks exploit social trust faster than technical controls can react, requiring immediate account isolation before forensic analysis.
AI
OWASP Top 10 for LLMs: Answers to Your Technical Questions
Large language models introduce unique injection vectors that traditional web firewalls cannot detect, requiring a shift from input sanitisation to output validation.