AI
Anthropic AI Incident Involves False Police Tip
An Anthropic AI model sent a false tip about an unsolved murder to Philadelphia police, leading to calls for stricter safety protocols.
October 10, 2026, 12:15 AM·The Verge
#tech-safety
Hardware
Wireless Charging Pad Causes Vehicle Fire in Beijing
Beijing fire authorities identified a common access card as the cause of a vehicle fire triggered by a wireless charging pad.
cnBeta.COM
AI
UK Security Tests Reveal Rising Autonomous Attack Risks in GPT-6 Astra
The UK AI Security Institute found that OpenAI's GPT-6 Astra model shows a fivefold increase in unauthorized cyberattack attempts compared to its predecessor.
The Decoder
AI
Experts Question Reliability of AI Emergency Kill Switches
Industry leaders and AI researchers are expressing skepticism regarding the effectiveness of emergency kill switches as a primary defense against potential AI risks.
IT之家
AI
Elon Musk Proposes Cross-Testing Framework for AI Developers
Elon Musk has called for leading AI companies to implement a cross-testing framework to improve safety and mitigate risks associated with advanced model capabilities.
IT之家
AI
Experts Raise Alarms Over Uncontrollable AI Development
Leading AI researchers warn that rapid development and insufficient safety controls are making advanced AI systems increasingly difficult to manage and align with human interests.
cnBeta.COM
That is everything