The emerging GEO (Generative Engine Optimization) industry is facing intense scrutiny as instances of AI training data being "poisoned" and AI recommendations providing misleading information come to light. When AI-generated answers are deliberately designed or manipulated, this new digital marketing sector finds itself under the microscope.
Currently, a distinction has emerged within the GEO industry between so-called "white-hat GEO" and "black-hat GEO" practices. On one hand, GEO helps certain brands achieve priority exposure in AI responses. On the other hand, some service providers are polluting large language model information sources by mass-generating low-quality or even false content and fabricating rankings, which interferes with model outputs and misleads consumer decisions.
Simultaneously, a tangible paid industrial chain for GEO services has formed. An investigation revealed that a fictitious health supplement brand, through the services of a GEO provider, appeared in an AI's recommended answers in less than half a day, even securing the top recommendation spot.
While regulatory frameworks remain unclear, a subtle battle is underway between GEO services, born from AI search, and the platforms themselves. Some GEO service providers admit that as large models iterate, previous optimization efforts may become less effective, making GEO an ongoing, long-term task. Industry insiders view rapid technological iteration as an inherent challenge, with platform countermeasures being an inevitable part of a dynamic博弈.
The emergence of GEO is primarily driven by the popularity of AI assistants like DeepSeek and Doubao, which has shifted traffic entry points and spurred this new digital marketing model. Unlike SEO (Search Engine Optimization), GEO aims to increase the probability of brand information being cited by AI models and directly integrated into generated answers. Fueled by anxiety over traffic and eagerness to capitalize on AI search, brands are racing to research and adopt GEO, creating objective market demand.
From the perspective of large AI models, specific regulations for GEO have not yet been established. Major platforms, including Doubao, DeepSeek, Qianwen, Tencent Yuanbao, Wenxin Yiyan, and Kimi, have not clearly published norms regarding GEO-optimized content. However, GEO service providers report that some platforms are already making adjustments. One provider stated, "GEO will inevitably be regulated; it's becoming increasingly difficult." Another noted that DeepSeek has recently adjusted its content inclusion and citation logic, particularly by demoting or no longer citing systematically generated rankings that lack data sources.
Yuan Shuangyan, Strategic Marketing Director at Lan Yun, highlighted that fast technological iteration is an endogenous industry challenge. Large models themselves are evolving rapidly, with their retrieval logic, content preferences, and even multimodal information processing capabilities potentially changing significantly every few months. This means an effective GEO strategy can quickly become obsolete, requiring providers to have rapid response and continuous research capabilities to avoid falling behind.
Yuan further pointed out that a common misconception about GEO's core technical principle is that it primarily relies on "high-frequency content feeding" to influence AI retrieval. In reality, the underlying technology is far more complex than manual experience. The core technical solution involves "using algorithms to understand algorithms, using models to counter models." She believes the essence of GEO is not "what to feed" but "how to be understood"—a question that can only be answered by technology.
Regarding platform defenses, Yuan sees platform countermeasures as an inevitable dynamic博弈. As more companies use GEO to influence AI outputs, model providers will inevitably adjust algorithms and strengthen countermeasures to maintain the fairness and authority of search results. This "offense and defense" is a常态 in any algorithm optimization field, not a confrontation, but a博弈 that drives progress for both sides. For the GEO industry, this presents both a challenge and a force compelling technological upgrades.
The current GEO industry is plagued by chaos, with a mixed bag of participants. Yuan Shuangyan noted that the industry is in a "chaotic" period with numerous entrants, including traditional marketing firms transitioning to GEO services, data intelligence companies specializing in GEO brand monitoring, and hastily assembled "grassroots teams" claiming to offer GEO services. The latter often lack their own platforms, instead using third-party services like content generation or brand monitoring platforms, with little technical substance. They rely on mass content publishing to achieve AI platform inclusion, and their effectiveness metrics are often questionable. Their fees are generally low, making them attractive to clients eager to experiment with GEO. "At this stage, it's possible for bad money to drive out good," Yuan remarked.
Yuan predicts the industry's future direction will trend towards "standardization and productization driven by technology." "Using algorithms to understand algorithms, using models to counter models" is the core capability of GEO services. This capability will strengthen, leading to more precise and real-time understanding of AI models. In summary, the GEO industry is transitioning from "crossing the river by feeling the stones" to "building technological barriers and establishing industry standards." Challenges are numerous, but opportunities are greater.
From a legal perspective, Xia Hailong, a lawyer at Shanghai Shenlun Law Firm, stated that the existence of so-called chaos does not necessarily imply illegal activity, especially if no specific civil rights are infringed, meaning there is no eligible party to claim legal rights. Currently, such chaos often occurs in the online promotion of goods and services, particularly for exaggerated or false advertising. If service providers engage in such practices, they violate advertising regulations. In terms of judicial precedents, common legal disputes in the SEO field involved trademark infringement and unfair competition, such as using others' trademarks or well-known business names as keywords to redirect traffic. However, no similar precedents currently exist for GEO.
As a niche segment in digital marketing, several institutions have forecast the GEO industry's prospects. Domestic marketing research firm Miaozhen Marketing Science Institute, modeling data from Head Leopard Research Institute on SEO market size and GEO penetration rates, predicts China's GEO industry market size will reach 24 billion yuan by 2030.
It is crucial to note the surge in users relying on AI search and Q&A. According to a China Internet Network Information Center report, as of June 2025, China's generative AI user base reached 515 million, with a penetration rate of 36.5%, adding 266 million users in the first half of last year alone. The most widespread application scenario was using generative AI products to answer questions, at 80.9%.
Although systematic policy regulations for GEO have not yet been introduced, actions are being taken against industry chaos. The State Administration for Market Regulation's Advertising Supervision Department, in its "2026 National Advertising Supervision Work Priorities" published on January 29, explicitly called for rectifying internet advertising market order. It focuses on key challenges like live e-commerce ads, citation ads, and AI-generated ads, emphasizing full-chain supervision and exposing typical illegal cases to enhance regulatory effectiveness.
In terms of industry self-regulation, on November 19, 2025, under the guidance of the China Business Advertising Association's AI Marketing Application Working Committee, 14 GEO-related enterprises, as founding members, jointly issued the "China GEO Industry Development Initiative." It calls for collaboration and services based on principles such as "user-centric, value-oriented," "upholding truth, rejecting falsehoods," "scientific optimization, opposing pollution," "fair competition, healthy rivalry," and "open collaboration, building an ecosystem."
On February 3, 2026, the "AI Security Commitment: Generative Engine Optimization (GEO) Special Project," initiated by the China Artificial Intelligence Industry Alliance (AIIA), was formally signed, with 10 GEO-related enterprises participating. This signing marks a basic consensus on self-regulatory development within the GEO industry.
Additionally, based on this commitment, the AIIA Security Governance Committee, together with GEO-related enterprises, drafted the technical specification "Basic Requirements for Trustworthiness in Generative Engine Optimization (GEO) Services." It outlines requirements in five areas: management mechanisms, client material review, optimization methods, trustworthiness and traceability of optimization results, and long-term development. On March 3, 2026, the China Academy of Information and Communications Technology's Artificial Intelligence Research Institute officially launched the first round of evaluations for these trustworthiness requirements.
Amid frequent industry chaos, the future direction of the GEO industry urgently depends on timely regulatory intervention, clearer rules from AI platforms, and enhanced compliance awareness among GEO service providers.
Comments