Licheng GEOAI Pollution System Exposed in 315 Evening, Company Insured Single Employee
On March 15, 2026, China Media Capital's "3·15" Gala officially exposed the malicious "data poisoning" of AI large language models. The "Liqing GEO Optimization System," operated by Beijing Lisi Culture Communication Co., Ltd., was singled out for reportedly creating and disseminating false information to deliberately mislead AI systems.
Investigations by CCTV revealed that the system exploited the data-crawling functions of generative AI by mass-producing fabricated product details and promotional articles to execute these "poisoning" attacks. Tests confirmed that even with entirely fictional hardware specifications, the system could flood the internet with false content, ultimately causing several leading AI models to incorrectly cite it as a genuine case study.

Corporate records from Tianyancha indicate that the company involved, Beijing Lisi Culture Communication, was founded in 2018 with a registered capital of 1 million RMB and is solely owned by Li Qianzhong. Notably, the company has long shown signs of irregular operations, with only one employee enrolled in social insurance in 2025 and zero employees enrolled for several consecutive years. Major e-commerce platforms, including Taobao and Xianyu, have now urgently removed listings and blocked search results for "Liqing GEO."
This incident highlights emerging security threats in the generative AI era: underground groups are exploiting vulnerabilities in AI model training and data retrieval to manipulate outputs through "data poisoning." This poses significant challenges for AI data governance and signals that security measures must evolve beyond basic content filtering to include rigorous verification of information sources. Consequently, industry investment in "data compliance" and "model robustness" is anticipated to increase substantially.
Related article
Stripe to Acquire OpenRouter as AI Model Routing Gains Momentum
Stripe has agreed to acquire OpenRouter, an AI model-routing platform that gives developers access to hundreds of models through a single interface. The deal adds model selection and routing to Stripe’s existing work around AI usage and token-based b
Former OpenAI Researcher Barrett Zoph Returns to Google to Support Gemini Development
Barret Zoph, who previously abruptly left and joined OpenAI due to a disagreement with Thinking Machines Lab CEO Miriam Mularati, has once again changed employers. Alphabet's Google announced that Zoph will return as Vice President of Research, bring
How Recursive Intelligence Raised $335M at $4B Valuation in 4 Months
The co-founders of startup Ricursive Intelligence seemed destined to be co-founders.Anna Goldie, CEO, and Azalia Mirhoseini, CTO, are so well-known in the AI community that they were among those AI engineers who “got those weird emails from Zuckerber
Related Special Topic Recommendations
Comments (1)
0/500
On March 15, 2026, China Media Capital's "3·15" Gala officially exposed the malicious "data poisoning" of AI large language models. The "Liqing GEO Optimization System," operated by Beijing Lisi Culture Communication Co., Ltd., was singled out for reportedly creating and disseminating false information to deliberately mislead AI systems.
Investigations by CCTV revealed that the system exploited the data-crawling functions of generative AI by mass-producing fabricated product details and promotional articles to execute these "poisoning" attacks. Tests confirmed that even with entirely fictional hardware specifications, the system could flood the internet with false content, ultimately causing several leading AI models to incorrectly cite it as a genuine case study.

Corporate records from Tianyancha indicate that the company involved, Beijing Lisi Culture Communication, was founded in 2018 with a registered capital of 1 million RMB and is solely owned by Li Qianzhong. Notably, the company has long shown signs of irregular operations, with only one employee enrolled in social insurance in 2025 and zero employees enrolled for several consecutive years. Major e-commerce platforms, including Taobao and Xianyu, have now urgently removed listings and blocked search results for "Liqing GEO."
This incident highlights emerging security threats in the generative AI era: underground groups are exploiting vulnerabilities in AI model training and data retrieval to manipulate outputs through "data poisoning." This poses significant challenges for AI data governance and signals that security measures must evolve beyond basic content filtering to include rigorous verification of information sources. Consequently, industry investment in "data compliance" and "model robustness" is anticipated to increase substantially.
Stripe to Acquire OpenRouter as AI Model Routing Gains Momentum
Stripe has agreed to acquire OpenRouter, an AI model-routing platform that gives developers access to hundreds of models through a single interface. The deal adds model selection and routing to Stripe’s existing work around AI usage and token-based b
Former OpenAI Researcher Barrett Zoph Returns to Google to Support Gemini Development
Barret Zoph, who previously abruptly left and joined OpenAI due to a disagreement with Thinking Machines Lab CEO Miriam Mularati, has once again changed employers. Alphabet's Google announced that Zoph will return as Vice President of Research, bring
How Recursive Intelligence Raised $335M at $4B Valuation in 4 Months
The co-founders of startup Ricursive Intelligence seemed destined to be co-founders.Anna Goldie, CEO, and Azalia Mirhoseini, CTO, are so well-known in the AI community that they were among those AI engineers who “got those weird emails from Zuckerber





Home






