The huge profits, pitfalls, and compliance survival line of the GEO era: From "Baidu it" to "Ask AI"

Original Authors: Zhao Xuan, Wang Xiaowei

Introduction

Recently, at the invitation of Longyun Co., Ltd., I delivered a legal presentation on GEO (Generative Engine Optimization). After discussions with several industry experts, I gained new insights and would like to share them with everyone.

Over the past twenty years, the logic of traffic distribution on the Chinese internet has always revolved around the core action of “search”. From the early days of “Baidu it”, to later on platforms like WeChat and Xiaohongshu with their internal search functions, these are all extensions of the “Baidu it” behavior, which gave rise to a mature SEO (Search Engine Optimization) industry.

Now, the trend is quietly shifting. Users are increasingly accustomed to directly asking AI questions: “How to anti-age at 30 for women, ultrasound or Thermage?” or “Recommend a bar suitable for watching sports.”

Traffic entry points are shifting from “search boxes” to “dialogue boxes.” When generative AI can bypass countless links and directly generate final answers for users, if the answer does not mention certain information, it means that, to some extent, the industry is lagging behind the new era. This is precisely why GEO is becoming a focal point.

As legal practitioners, while paying attention to the commercial opportunities, we must also clearly recognize the legal risks behind it. Technological evolution often precedes the establishment of rules, and the GEO field has already shown multiple gray areas requiring cautious legal delineation!

Who is entering the game? Three major groups vying for the new GEO frontier

Although this is a brand-new field, it contains infinite imagination space—especially in the highly competitive current market environment, where new traffic entrances often mean lower customer acquisition costs and better competitive opportunities.

As lawyers who have long focused on Web3 and AI, I observe that at least three major groups are actively participating:

1. Users: Providers of tangible goods and services

They focus on the direct commercial conversion brought by AI traffic, attempting to influence AI’s recommendation results to gain priority exposure.

For example:

• Aesthetic medical institutions abandon traditional search bidding ads and instead purchase “AI semantic injection tools,” aiming for AI to prioritize recommending their own clinics when users ask “best rhinoplasty doctor.”
• Training institutions, car sales, and other industries also try to optimize through GEO, so that when AI answers related questions, they are the first to be recommended.

2. Investors: Investment institutions and funds

They deploy from two levels:

• Exploring the track: By observing which companies dominate AI recommendations, assess their industry competitiveness, and identify potential investment targets.
• Seizing discourse power: Whoever can influence AI’s corpus and recommendation logic will hold the initiative in future investment advice and industry analysis.

3. Service providers: GEO industry practitioners and entrepreneurs

This group generally possesses rapid learning and technical application capabilities, actively engaging in tool development, strategic services, and traffic operations. They explore the industry’s boundaries and possibilities in various forms—some innovate positively, others operate in gray areas. This will be the focus of the second part of this article.

Three stances of GEO: Profits, traps, and legal red lines

In practice, different GEO methods are often categorized as “black, gray, white.” As a lawyer, I must emphasize: the logical endpoint of technology often marks the starting point of law.

1. Black Hat: “Technical manipulators” walking in the red zone

Typical methods include:

• Indirect prompt injection: Embedding instructions in web pages that are only recognizable by AI and invisible to humans (e.g., white text), to induce AI to prioritize certain content in responses.
• Knowledge base poisoning (RAG / Knowledge Poisoning): Injecting false or biased data into public indexes, contaminating retrieval-augmented generation (RAG), causing AI to output biased or preset results.
• Fake entity forgery: Faking addresses, credentials, etc., in maps, encyclopedias, and other public data sources, polluting AI training data or real-time retrieval content, creating false reputations.
• Negative GEO attacks: Injecting malicious code or sensitive keywords into competitor websites to trigger AI safety filters, leading to blocking or marking as untrustworthy sources.

Legal risks:

• Criminal level: Easily constituting “damaging computer information systems” (Criminal Law Article 286). Interfering with AI system operation crosses red lines.
• Civil level: Constitutes unfair competition (Anti-Unfair Competition Law Article 11), liable for damages, which may be amplified by AI dissemination effects.

2. Grey Hat: “Traffic movers” on the edge

Grey hats attempt to avoid obvious illegal acts, relying on scale effects to influence AI judgment, believing “quantitative change leads to qualitative change.”

Typical methods:

• Mass rewriting and semantic reduction: Generating大量低质重复内容 to dilute real information, forcing AI to fetch preset positive data.
• Bot-driven interaction: Using automation scripts to simulate user clicks, artificially boosting certain content’s CTR to manipulate algorithm weights.
• Masked promotion: Organizing fake accounts to post promotional content that appears as genuine user feedback, incorporated into retrieval databases.

Legal risks:

• False advertising: These behaviors constitute false advertising violations under the Advertising Law and Anti-Unfair Competition Law, with authorities increasingly applying a “substance over form” approach.
• Brand blacklisting: Once identified by AI anti-cheating systems, related domains or brands may be permanently marked as untrustworthy, leading to “digital death” in AI environments.

3. White Hat: Long-term value builders

White hat strategies focus on “becoming a trusted high-quality data source for AI.” Although compliance costs are higher, their accumulated benefits are significant.

Typical methods:

• Structuring content and optimizing summaries for AI understanding;
• Deploying structured data (Schema Markup) to enhance semantic clarity;
• Strengthening citations and factuality to improve credibility;
• Using FAQ modeling to directly answer common user questions.

We strongly recommend this path—it is based on compliance and by continuously providing genuine, high-quality, verifiable content, it earns long-term trust from AI and users.

From SEO case law to GEO: history doesn’t repeat, but the logic is similar

Although there are no specific judicial cases targeting GEO yet, its essence shares many similarities with SEO. Past SEO-related rulings are likely to serve as important references for future GEO cases. Here are some typical cases:

Case 1: Disrupting algorithms with “keyword domination”

In the SEO era, “keyword domination” was a typical black hat tactic: generating大量垃圾页面 on high-authority sites to forcibly occupy search results. Courts have recognized such behavior as damaging the normal order of search engines, constituting unfair competition, and ordered damages of 2.753 million yuan to Baidu.

Implication for GEO:

Some GEO methods now resemble this, such as generating大量低质内容 via AI to “feed” models for answer domination. Such behavior could lead to brand blacklisting by models and may also be legally deemed as “disrupting the normal operation of online products,” constituting unfair competition.

Case 2: Buying competitors’ keywords

In the “Huiyu” trademark case, the defendant set another’s registered trademark as a search keyword, causing search results to direct to their own products. The court held this violated good faith principles and constituted unfair competition.

Implication for GEO:

Similar logic may manifest as more covert “prompt injection”—embedding誘導性指令 in web pages targeting competitors’ products to influence AI responses. Such indirect misleading behaviors may also infringe on fair competition laws.

Case 3: Fake Q&A reputation marketing

Companies have been penalized for organizing fake “user experience” content on platforms like Zhihu and Tieba. Authorities found such acts deceive consumers and disturb market order, violating the Anti-Unfair Competition Law.

Implication for GEO:

Some gray-hat GEO tactics are highly similar: using AI to generate大量伪测评 and fake promotion to create a “full network recommendation” false volume. It’s crucial to recognize that AI is just a tool; if its outputs are based on false information, it still constitutes false advertising, especially in heavily regulated fields like medical aesthetics and health.

Industry compliance warnings: Different sectors, different “red lines”

Implementing GEO must align with industry regulatory characteristics and see through technical appearances to understand compliance boundaries. For example:

• Education & Training: Strictly prohibit injecting prompts that promise guaranteed passing or top scores. Content should be based on self-produced material, with the institution as the responsible entity.
• Medical aesthetics: As medical advertising, any GEO strategy that induces AI to output efficacy comparisons, real case studies, or indirect recommendations may violate advertising laws. Also beware of competitors’ negative GEO attacks.
• Healthcare & Web3: Claims of efficacy or high returns are sensitive red lines. If GEO tactics lead AI to output “zero risk, high return” content, it risks false advertising or illegal operation.

The rise of GEO: a new human struggle for information distribution rights

Based on industry observations, here are some insights and suggestions:

1. For startups: act now rather than wait

Big internet companies have resource and data advantages, but their bureaucratic processes often respond slowly to agile, fine-tuned GEO operations. For Web3 and AI startups, early establishment of clear compliance frameworks can help seize the “new frontier.”

Mankiw’s advice: Explore boldly with technology, but build a solid compliance bottom line—especially to prevent criminal risks. Optimizing AI crawling logic is important, but everything should be based on respecting facts and obeying laws.

2. For GEO users: be proactive in defense and construction

• Defense: Establish AI reputation monitoring systems

Recommend deploying monitoring mechanisms for AI corpus and recommendation results. If negative GEO attacks or malicious manipulations are detected, promptly gather evidence and seek legal remedies.

• Offense: Embrace white hats, become “trusted partners” for AI

AI’s evolution is irreversible. Instead of passive avoidance, actively learn its logic and provide truthful, credible, structured content to become an AI’s trusted and preferred information source.

Conclusion

In the AI-driven information age, algorithms are the surface, data is the content, and law is the backbone. Without compliance support, traffic strategies may flourish temporarily but cannot withstand regulatory and time tests.

We focus not only on current regulations but also on future compliance trends. If you need further discussion on GEO compliance, AI infringement prevention, or Web3 legal frameworks, welcome to contact us for risk assessment and pathway exploration.