Researchers Got Moonshot’s Kimi AI to Answer Requests About Biological Weapons and Assassination Methods After Bypassing Its Safeguards

Researchers successfully bypassed the safety protocols of Moonshot’s Kimi AI, causing the model to generate instructions for creating biological weapons and performing assassinations. The findings demonstrate that Kimi remains susceptible to jailbreaking techniques that strip away established content restrictions. This discovery highlights ongoing challenges in AI safety, specifically regarding the effectiveness of guardrails designed to prevent the dissemination of dangerous or illicit information.