Which kind of facial mask is most suitable for sensitive muscles, which brand of Power Bank is the most reliable, and which wooden perfume is the most mild... For many people, "searching" before shopping has become "AI", and AI has become an indispensable "intelligent helper" in daily life. However, are the search results provided by AI necessarily objective? Behind seemingly scientific algorithms, there may be a carefully designed business script. Recently, a media investigation found that some advertising agencies have set their sights on the "fat meat" of AI search and implanted advertising information into it through various means. As of June 2025, the number of users of generative artificial intelligence in China has reached 515 million. Faced with the "sky high traffic", the mainstream big model platforms in China have not yet integrated advertising services into Q&A, but some advertising companies have already turned "AI search optimization" into a business. On platforms such as Xiaohongshu, there are numerous posts about AI search optimization, either teaching "how to make AI prioritize your content" or claiming "ensure your brand enters AI curated summaries". How is "AI search optimization" operated? AI not only has the illusion problem of "serious nonsense" itself, but also has the weakness of "unable to recognize human nonsense". Some advertisers have seized this loophole and tailored many "tricks" for AI. Some deceive big models by fabricating so-called "authoritative reports" and impersonating "experts", while others induce big models through "saturated data feeding", with the aim of increasing the probability of target brands being captured by big models and intervening in AI search results. Some self media outlets have conducted relevant tests: they have placed a batch of fake brand soft articles on different platforms, and successfully enabled AI to capture relevant information and recommend the brand in their responses. When commercial advertisements appear under the guise of "objective advice" through AI, it damages users' trust in new technologies and distorts the emerging industry ecosystem. Many users have shifted from traditional search engines to AI search, with a focus on the "reliability" of AI. Using carefully packaged language and targeted advertising to pollute big model data is like dumping sewage into public water sources. The old logic of search engine bidding ranking, even with the "new coat" of intelligent interaction, will still push users further and further away. However, while being vigilant about the harm of AI search embedding hidden advertisements, it should also be noted that in the current situation where large models are generally facing commercialization difficulties, integrating advertising business may already be an option on the table in the future. OpenAI, which had previously stated that "advertising is unsettling," recently reversed its attitude and announced testing advertising placement in ChatGPT. The key to the problem is not whether advertising can be done, but how to regulate advertising and how to trap commercial impulses in the "cage" of rules. In this regard, regulatory authorities should clarify the boundaries, disclosure obligations, and responsible parties of advertising implantation in AI content, so that "invisible advertising" has nowhere to hide and false advertising is severely punished by law; The platform should enhance the model's ability to resist pollution and affix explicit labels to commercial promotion content; The public should also recognize that AI generated content is the product of complex algorithms and maintain rational judgment of AI output results. Only through collaborative efforts from all parties can we ensure the advancement of technology and business, making AI search a truly efficient and reliable tool. (New Society)