❌

Normal view

Before yesterdayMain stream

Feed: Artificial Intelligence Latest
DeepSeek’s Safety Guardrails Failed Every Test Researchers Threw at Its AI Chatbot 31 January 2025 at 12:30

DeepSeek’s Safety Guardrails Failed Every Test Researchers Threw at Its AI Chatbot

Feed: Artificial Intelligence Latest

By:Matt Burgess， Lily Hay Newman

31 January 2025 at 12:30

Security researchers tested 50 well-known jailbreaks against DeepSeek’s popular new AI chatbot. It didn’t stop a single one.