The China Consumers Association today issued a consumer advisory on artificial intelligence services, reminding consumers to understand rationally the informational assistance and reference value of general-purpose generative AI. For important consumer information involving prices and fees, contract terms, after-sales services, legal liability, and other matters, consumers should not make decisions solely on the basis of AI-generated content. They should verify and confirm such information through the business’s official channels, contract documents, regulatory authorities, and other authoritative sources.
The full text follows:
At present, generative AI technology centered on large language models is rapidly becoming integrated into various aspects of consumer life. While AI improves consumer convenience and service efficiency, the inherent uncertainty of information generation may also result in inaccurate information, imprecise wording, or discrepancies between promises and actual circumstances. Examples include providing fee information that does not match reality, making compensation statements that are difficult to fulfill, and offering one-sided or insufficiently cautious legal advice. Such issues may affect consumers’ judgment and choices and may even cause financial losses.
According to relevant public feedback and recent complaints received by consumer protection organizations nationwide, consumer problems arising from AI services have different characteristics in different application scenarios. The AI that consumers encounter in daily life can generally be divided into two categories. One is general-purpose generative AI services for the public, which mainly generate responses based on public information, model training data, and user questions and generally do not represent the official views of a particular business. The other is AI customer service tools deployed and used by businesses, which directly participate in business service processes such as product introductions, fee explanations, promotion explanations, and after-sales handling.
General-purpose generative AI services for the public are not spokespersons for specific businesses. Their generated content generally cannot be treated directly as a formal commitment by a platform or merchant, nor can it replace contractual terms or the opinions of authoritative institutions. Their primary role is to provide informational assistance and reference. In practice, some general-purpose generative AI systems may make unsupported guarantees covering losses suffered by consumers during consumer consultations or rights-protection assistance, make overly definitive judgments about litigation outcomes, provide incorrect legal guidance on litigation jurisdiction, or even generate false information about nonexistent payments, deposits, or case processing. If consumers lack the ability to properly distinguish such generated content and fail to verify it through a business’s official channels, contract documents, or relevant authoritative sources, they may make incorrect decisions about refunds, rights protection, or litigation. This can affect normal consumption decisions and lawful rights protection, and may also cause financial losses. The Interim Measures for the Management of Generative Artificial Intelligence Services expressly provide that AI service providers should, based on the characteristics of their service types, take effective measures to improve the transparency of generative AI services and the accuracy and reliability of generated content. They should clearly identify and disclose the intended users, situations, and purposes of their services, and guide users to understand and use generative AI technology scientifically, rationally, and in accordance with the law.
The China Consumers Association recommends that providers of general-purpose generative AI services for the public further strengthen risk warnings for key scenarios and information quality management. In scenarios closely related to consumers’ important rights and interests, such as financial services, daily-life services, and rights-protection decisions, providers should appropriately explain the limitations of AI-generated content and remind consumers to further verify important information. At the same time, they should continuously improve model accuracy, enhance content review, risk warning, and error-correction mechanisms, improve response quality, minimize inaccurate information and content that could easily cause misunderstandings, and reduce the risk of consumers suffering losses due to excessive reliance on AI-generated information.
AI customer service tools independently deployed by businesses are part of their business service chains. In practice, some AI customer service systems may provide price and fee information that is inconsistent with reality, inaccurate explanations of promotional rules, or insufficient notices of important restrictions. After a dispute arises, some businesses may also attempt to sever responsibility for relevant responses or evade their responsibilities on the grounds that they were “automatically generated by AI” or were “automatic system replies.” These problems can easily cause consumers to misunderstand transaction prices, fee structures, and after-sales services, affecting their rights to know and to fair transactions. The Consumer Rights Protection Law provides that consumers have the right to know the true circumstances of the goods they purchase or use and the services they receive. When consumers ask businesses questions about the quality or usage of the goods or services they provide, the businesses must give truthful and clear answers.
The China Consumers Association urges businesses that deploy AI customer service tools to strengthen review and management, ensuring that information provided by AI customer service is consistent with the actual goods and services, prices and fees, promotional rules, contractual terms, and after-sales policies. Businesses should also provide convenient and effective human customer service and dispute-resolution channels. While using AI to improve business and service efficiency, businesses should establish corresponding responsibility mechanisms. For important responses concerning prices, fees, after-sales services, and other matters that AI customer service provides to consumers during business operations and service delivery, businesses may not simply claim exemption from liability on the grounds that the responses were “automatically generated by an algorithm,” and may not unreasonably reduce or exempt themselves from legally mandated responsibilities through relevant disclaimers.
The China Consumers Association calls on relevant authorities to further improve regulations governing the application of AI in the consumer sector. Taking into account different service types and application scenarios, authorities should further clarify the boundaries of responsibility and governance requirements for service providers, businesses, and other parties, promoting a clearly defined and orderly AI consumer service environment. Consumers are also reminded to understand rationally the informational assistance and reference value of general-purpose generative AI. For important consumer information involving prices and fees, contract terms, after-sales services, legal liability, and other matters, consumers should not make decisions solely on the basis of AI-generated content. They should verify and confirm such information through the business’s official channels, contract documents, regulatory authorities, and other authoritative sources.