Artificial intelligence relies heavily on machine learning techniques for its intelligence. Through the training of data, the design of algorithms and the operation of models, machine learning technology is able to transform problems into mathematical and statistical solutions, thus giving the logic of artificial intelligence technology a mathematical logic. Artificial intelligence technology generates the phenomenon of “consciousness” on the basis of socialized and collaborative mental production, and generates the ideology that corresponds to it, driven by the logic of capital application and technology abuse. The ideology based on the new characteristics, new laws and technological logic of AI technology is also facing the possibility of new ideological risks brought about by the provisions of AI technology. Ideological risks in the era of artificial intelligence in the field of ideology generation, dissemination and identity respectively present “capital logic, algorithmic black box and algorithmic ethics”, “algorithmic recommendation, filtering bubbles and information cocoon”, and “subcultural space, value penetration and digital fetishism”. The different stages of risk are inherited and intertwined in reality, constructing a complex ideological risk network. Based on a comprehensive grasp of the logic of the generation of ideological risks in the AI era and a systematic sorting out of risk phenomena, we can target the root causes of ideological risks in the AI era through the application of the logic of socialism, the technical prevention of ideological risks in the AI era through the advancement of systematic regulation and technological innovation, and the systematic prevention of ideological risks in the AI era through the creation of a virtuous ecology and other pluralistic governance. This is also the ideological attempt of Chinese modernization to create a digital civilization in the age of artificial intelligence.