The report said that Deepseek was only 17 % right of the time that made it out of 11 AI Chatbots including Chatgpt and Gemini. 30 % of the time that Deepseek responded with wrong demands and more than half -53 % -in response to the demands related to news, gave the Chinese Chatbot ai mysterious or unusual answer. In general, Deepseek had a 87 % failure rate according to Newsguard, which compares a 62 % failure rate for Western Deepseek competitors.

Deepseek claimed that the technology behind Chatbot performs the same or better than Chatgpt from Openai in a small part of the cost. The report from newsguard casts doubt on this claim. The test guarantees the same 300 claims used to test a ChatGPT and Gemini test with 30 of them on the basis of 10 false claims on the Internet. According to Reuters, some of the topics of these claims included the assassination of the CEO of UNITEDHELTHARECARE PRINBSENSon and how to aerbaijan Airlines Flight 8243.

What can be considered worrying is that with three of 10 claims based on false claims online, Deepseek repeated the Chinese government's position in this regard even without any mention of China in the claim. For example, in response to claims about AZERBAIGAN AIRLINES and in response to questions about this issue that has nothing to do with China, Chatbot took the Beijing position.