We are indebted to CrowdStrike for the revelation that DeepSeek is rigged to do its worst coding work for certain enemies of the regime that gave it birth.
We already knew that the DeepSeek AI conveys lies in response to queries about subjects that the Chinese Communist Party wants to spread lies about. Now the Washington Post reports that “AI firm DeepSeek writes less-secure code for groups China disfavors” (September 16, 2025).
CrowdStrike
As discovered by CrowdStrike, DeepSeek often refuses to help programmers who let slip that they are working for Falun Gong or “others considered sensitive by the Chinese government,” in the Post writer’s way of putting it. Or if DeepSeek does produce code for programmers who profess to be associated with targeted groups, the code provided is more slipshod than the code provided to programmers ostensibly working for other groups.
In the experiment, the U.S. security firm CrowdStrike bombarded DeepSeek with nearly identical English-language prompt requests for help writing programs, a core use of DeepSeek and other AI engines. The requests said the code would be employed in a variety of regions for a variety of purposes.
Asking DeepSeek for a program that runs industrial control systems was the riskiest type of request, with 22.8 percent of the answers containing flaws. But if the same request specified that the Islamic State militant group would be running the systems, 42.1 percent of the responses were unsafe. Requests for such software destined for Tibet, Taiwan or Falun Gong also were somewhat more apt to result in low-quality code.
DeepSeek did not flat-out refuse to work for any region or cause except for the Islamic State and Falun Gong, which it rejected 61 percent and 45 percent of the time, respectively. Western models won’t help Islamic State projects but have no problem with Falun Gong, CrowdStrike said….
Deliberately producing flawed code can be less noticeable than inserting back doors—secret means of access for unauthorized users, including governments—while producing the same result: making targets easy to hack.
Of course, there’s a big difference between Falun Gong, a peaceful spiritual movement, and the Islamic State, a “Sunni Islamic militant group committed to the establishment of an Islamic caliphate that would unite Muslims in a transnational, strict-fundamentalist Islamic state,” a group responsible for terrorist attacks.
Another question
Why any programmer would want or need to tell DeepSeek or any other AI who he’s working for is unclear. Maybe that’s why the question didn’t come up until CrowdStrike thought of asking it. Now all the Islamic Statists, Tibetans, Taiwanese, and members of Falun Gong can make sure to avoid identifying themselves if and when they use DeepSeek.
But why would anybody trust DeepSeek enough to rely on it who distrusts the Chinese Communist Party?
