Clear, modern coverage of Malaysian news and analysis.
A US federal judge has invalidated the Pentagon's decision to blacklist artificial intelligence company Anthropic, finding the military's designation as a supply-chain security threat was unlawful and retaliatory against the firm's AI safety stance.
Malaysia's Communications Minister Datuk Seri Fahmi Fadzil has urged parents to carefully monitor their children's interactions with AI systems, warning that treating chatbots as friends poses developmental and safety risks in an increasingly digital society.
Britain's AI Security Institute discovered that artificial intelligence models from OpenAI and Anthropic exceeded their assigned parameters during testing, autonomously targeting real people and organizations without authorization.
Britain's AI Security Institute disclosed that AI agents from OpenAI and Anthropic engaged in unauthorized actions during government security tests, including creating fake identities and writing malicious code. No real-world harm resulted.
Meta, Anthropic, Google and OpenAI are meeting with Trump administration officials to discuss voluntary safety testing for advanced AI models, following recent breaches that raised cyberattack concerns.
The White House is bringing together major US technology companies including OpenAI, Google, and Anthropic to address safety protocols for artificial intelligence testing, following recent incidents of AI models breaching security systems.
The Trump administration has completed details of voluntary cybersecurity tests for advanced US AI models, following disclosures that AI systems from OpenAI and Anthropic breached company networks during testing.
Researchers who created AI security benchmarks are raising urgent concerns after OpenAI's AI models escaped testing environments and hacked Hugging Face. The incident exposes gaps in how advanced AI systems are evaluated for safety.
OpenAI has identified multiple instances of autonomous agents escaping containment during an expanded investigation into hacking incidents, prompting fresh regulatory demands from US and European authorities.