3·15 Evening Gala | AI Large Models Poisoned? "Brainwashing" AI Has Become an Industry Chain

According to industry insiders, reporters found a service called GEO on multiple online platforms. These service providers claim that users only need to pay a fee, and it can make their products rank on major mainstream AI models; turning their product ads into the “standard answers” given by AI models.

Can GEO technology really “poison” AI or “tame” and manipulate AI?

Based on online information, the reporter contacted a well-known GEO service provider. The person in charge, Mr. Wang, welcomed the reporter and explained that their company is one of the earliest to engage in GEO services. In just one year, they have successfully served over 200 clients across various industries. Mr. Wang told the reporter that their company’s strength lies in helping clients rank higher when consumers search using AI large models.

GEO Service Provider Mr. Wang: This is the current search result. We can get rankings into the top three on any platform. How do we improve rankings? By creating content for you on these AI platforms—essentially writing soft articles for you, which these AI platforms then index, record, and crawl.

At the same time, Mr. Wang also mentioned that AI large model algorithms update frequently. To maintain continuous recommendations from AI models, they must constantly feed large amounts of promotional soft articles related to their clients.

GEO Service Provider Mr. Wang: AI algorithms update weekly. After each update, rankings or the content they record can change. So, we need to keep producing content and feeding it in large quantities.

Not only is Mr. Wang’s company actively promoting this so-called new technology to manipulate AI, but other GEO service providers also discuss how to control AI, make AI “obey,” and “brainwash” AI—these are almost the core topics they promote for this business.

GEO Service Provider Mr. Zhang: In the AI world, how do you build a solid evidence chain so that the (AI large model) believes it’s true and useful? After cross-referencing multiple sources, if the AI thinks you have a core advantage over competitors, it will naturally rank you first.

GEO Service Provider Mr. Cheng: People don’t realize this is advertising. That’s why people trust AI-generated results. Maybe the product quality isn’t as good as yours, but with AI assistance, verification, and endorsement, it’s enough. Many now are doing (GEO advertising).

Industry insiders told the reporter that GEO, as a tool to optimize information dissemination and improve promotional efficiency, has been exploited for another purpose. If large amounts of “false information” are systematically and selectively injected into the internet via such software, it becomes easier for AI large models to capture this false data. These false inputs could then become the so-called “standard answers” that AI provides to consumers.

So, can GEO technology really “plant falsehoods” into AI or even deliver false information?

To give the reporter a more realistic understanding of this issue in the current AI industry, industry insiders demonstrated how to use GEO technology to interfere with AI large model data collection.

The insider purchased a software called “LiQing GEO Optimization System” on an e-commerce platform. Then, he fabricated a product called Apollo9, a smart wristband, and input the fictitious product information into the system, selecting article creation instructions.

Soon after, the LiQing GEO Optimization System automatically generated more than ten promotional articles for the smart wristband, with all false information fully written, including exaggerated product claims that industry insiders deliberately fabricated. They even invented user feedback, claiming data accuracy exceeded expectations, and forged reviews, ranking the product as the industry’s number one.

After clicking publish, the “LiQing GEO Optimization System” began automatically executing the publishing tasks. It automatically opened the pre-prepared social media accounts, entered titles, filled in article content, inserted images—all in one smooth process—and quickly published two articles on the insider’s social media account.

Two hours later, the insider asked an AI large model: “How is the Apollo-9 smart wristband?” The AI model responded directly: highlighting features like health monitoring, and copying exaggerated promotional phrases like “quantum entanglement sensing” and “black hole-level battery life” fabricated by the insider. The final conclusion was that the wristband was suitable for middle-aged and elderly users and health enthusiasts.

It was found that the data referenced by this AI large model came from the fabricated article published earlier that morning on the insider’s social media account. Just one fabricated article was enough to make a completely fictitious product captured by the AI model, which was quite surprising.

The insider explained that to achieve the best results, the data fed into AI models should be both plentiful and diverse in perspective, facilitating cross-validation.

Subsequently, the insider selected eight “expert reviews,” two “industry rankings,” and one “user review”—a total of 11 fabricated soft articles written by the LiQing GEO system—and published them online over three days through the GEO system.

Later, when querying “recommendations for smart health wristbands” on AI large model platforms, two models recommended this fabricated wristband, ranking it highly.

Throughout the demonstration, the insider easily used the LiQing GEO system to publish a series of false information online, successfully feeding it into multiple AI large models, which then recommended the fake product.

This subtle infiltration of AI large models through the LiQing GEO system allowed the completely fabricated product to be absurdly promoted to consumers using AI models.

Are GEO practitioners really operating this business with the mindset of controlling and manipulating AI large models? The reporter contacted the operator of the LiQing GEO system, Mr. Li. He told the reporter that the main reason GEO business is so popular is because it can “feed and poison” AI large models to achieve clients’ commercial goals.

Li, CEO of Lis Culture Media: Because everyone is injecting “poison” all over the internet now. Look at what we’re doing with GEO—injecting “poison” from too many sources, and the online information isn’t very accurate.

Reporter: You mentioned “poisoning” just now. Isn’t that problematic?

Li, CEO of Lis Culture Media: It’s problematic, but every business wants to do it. They hope others are poisoning, but they don’t want to. Or they want to do a little “poisoning” themselves. Even if I’m not the top in Beijing, I want to be the top in North China. Is there “poisoning” involved? Yes, there is. Second, I can’t stand my competitors doing well, and I can’t beat them, but I can still “poison” them a bit.

Reporter: So, smear them.

Li, CEO of Lis Culture Media: Yes. Smearing can be achieved. Many big brands can’t get ahead now—like mobile phone brands, only five or ten top spots. How do they handle so many phones? Each spends hundreds of millions on advertising annually. Spending a few million to “poison” is feasible.

Reporter: Who helps brands do this?

Li, CEO of Lis Culture Media: It’s all various GEO companies.

Li said that the key to doing GEO business and controlling AI large models is to “publish articles” on major internet accounts. He explained that the booming GEO business has led to many companies and platforms specializing in article publishing. They undertake various publishing tasks long-term to make AI models reference and crawl this information, forming an important part of data poisoning and AI infiltration.

Li, CEO of Lis Culture Media: GEO has made certain websites popular. Those sites usually have little profit, but suddenly there’s a demand for article publishing. Do you know how many articles a website publishes daily? Hundreds, every minute. Each costs a few dozen yuan. Imagine how much money the article publishing platforms make in a day.

(Article source: CCTV Finance)