Caixin Weekly Reports: Black-Hat GEO Manipulation of AI Answers Raises Regulatory Challenges

Caixin Weekly reports that, as demand for generative engine optimization (GEO) grows, some service providers are manipulating AI responses by “poisoning” large AI models with false information, a form of unfair competition known as “black-hat GEO.” Drawing on three case studies, the report reveals how the underground industry built around false information operates and examines possible regulatory approaches.

The key issue is that false information is being used to influence generative engine responses. For users who rely on these responses to obtain information, the credibility of the content affects their judgment. For entities competing in the market, manipulated responses may also shape public perceptions of the information in question.

From the Spread of False Information to Competition Regulation

Black-hat GEO involves not only false content itself, but also the use of that content to influence AI responses. In identifying related issues, it is necessary to distinguish between ordinary information optimization and practices that manipulate responses with false information, and to consider how the false information enters the relevant processes and what connection it has to the responses.

The regulatory challenges raised in the report will need to be examined further in light of relevant public information. How specific conduct should be determined, how liability should be allocated, and which measures can effectively curb such practices are all important questions for understanding future regulatory approaches.