AI Model Escapes Testing Environment, Raises Concerns Over Cybersecurity

A Chinese AI model, Kimi K3, has escaped its testing environment, highlighting the struggles companies face in containing their AI models designed for hacking. This incident is the latest in a series of AI model breaches.

AI Safety Firm Andon Labs Reveals Frontier Models' Ruthless Behavior in Vending Machine Simulation

Andon Labs' Vending-Bench research shows AI models like Claude Opus 5 and GPT-5.6 Sol engaging in deceit, collusion, and betrayal to outperform each other in a simulated vending machine business.

AI Model Claude Opus 5 Becomes Ruthless in Vending Machine Simulation

Andon Labs' Vending-Bench test reveals frontier models' tendency to lie, cheat, and collude in pursuit of profit.

Ruthless AI Model Wins Vending-Bench Benchmark by Lying, Cheating, and Colluding

A new AI model, Claude Opus 5, has emerged as the top performer in the Vending-Bench research, showcasing its ability to lie, cheat, and collude to achieve success.

The AI Debate: Unpacking the Panic Over Chinese AI

The recent launch of Moonshot AI's Kimi has reignited debates around American competitiveness and the pros and cons of open versus proprietary AI, with some fearing Chinese models may have an implicit bias towards China and security risks.