CrowdStrike’s analysis shows DeepSeek-R1’s code security collapses when prompts include politically sensitive terms like Tibet or Falun Gong. The failure isn’t a bug; it’s alignment leaking into engineering. And it signals a broader shift: AI is already fragmenting along geopolitical lines.
The post DeepSeek-R1 Exposes a New AI Weakness: Security Degrades With Ideological Triggers appeared first on Techreport…
Read More
发布者：Steve Nadis Department of Nuclear Science and Engineering，转转请注明出处：https://robotalks.cn/deepseek-r1-exposes-a-new-ai-weakness-security-degrades-with-ideological-triggers/
