AI语料"投毒"产业链揭秘："用模型对抗模型" 百亿市场该如何健康发展？
2026-03-14
东方财富网
The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
In China, marketing firms exploit Generative Engine Optimization (GEO) to poison AI training data, causing large language models to recommend fictitious or low-quality products and services. This manipulation misleads consumers and distorts market information, with a paid industry emerging around influencing AI-generated answers and recommendations.[AI generated]
Why's our monitor labelling this an incident or hazard?
The event involves AI systems explicitly, as GEO practices manipulate AI large language models and AI-generated search results. The use of AI-generated false or misleading content that appears in AI outputs and influences consumer decisions constitutes harm to communities and consumers, fulfilling the criteria for harm under the AI Incident definition. The article documents realized harms (misleading AI answers, false brand recommendations) caused by AI system manipulation. Although regulatory and self-regulatory responses are underway, the primary focus is on the existing harms caused by AI system misuse and data poisoning. Therefore, this is an AI Incident rather than a hazard or complementary information.[AI generated]