In a research project called “Algorithmic Self-Portrait” initiated by the Russian Ministry of Internal Affairs, experts discovered that OpenAI violated its own data privacy policy and retained users' personal information to construct their digital portraits. Using this data, users can identify themselves even if they use a pseudonym, and there is also a risk of being manipulated — after adding false instructions to the documents processed by AI, AI will have a hidden influence on the decisions of the relevant personnel based on the characteristics of the person concerned, and then act on the judgment of public officials and scientists. To prevent risks, the Ministry of Internal Affairs advises users to check the memory and personalized settings of AI services, not upload work and personal information to them, de-identify requests, and use temporary chat windows.

In a research project called “Algorithmic Self-Portrait” initiated by the Russian Ministry of Internal Affairs, experts discovered that OpenAI violated its own data privacy policy and retained users' personal information to construct their digital portraits. Using this data, users can identify themselves even if they use a pseudonym, and there is also a risk of being manipulated — after adding false instructions to the documents processed by AI, AI will have a hidden influence on the decisions of the relevant personnel based on the characteristics of the person concerned, and then act on the judgment of public officials and scientists. To prevent risks, the Ministry of Internal Affairs advises users to check the memory and personalized settings of AI services, not upload work and personal information to them, de-identify requests, and use temporary chat windows.