Cisco researchers found it was much easier to trick DeepSeek into providing potentially harmful information compared to its rivals, such as ChatGPT, Google's Gemini, or Anthropic's Claude.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results
Feedback