As generative AI, driven by large language models, rapidly integrates into everyday consumer scenarios, it brings greater convenience and efficiency to shopping and services. Yet, the inherently uncertain nature of AI-generated information can also lead to inaccuracies, imprecise statements, and promises that fail to align with actual conditions. For instance, AI might present incorrect fee structures, offer compensation pledges that are hard to fulfill, or supply one-sided or inadequately considered legal guidance, all of which can skew consumer judgment and decision-making, potentially resulting in financial harm.

The China Consumers Association (CCA) reports that, based on public sentiment and recent complaints received by consumer councils nationwide, the consumer issues triggered by AI services exhibit distinct traits across different applications. The AI that consumers typically encounter falls into two main categories: one is general-purpose generative AI services open to the public, which primarily craft responses from publicly available information, model training data, and user queries, and generally do not represent the official stance of any particular business. The other is AI customer service tools deployed by businesses, which actively participate in core operational stages such as product introductions, cost explanations, activity clarifications, and after-sales handling.

The CCA advises that providers of public-facing, general-purpose generative AI services should enhance risk notifications and information quality controls in key scenarios. In areas that closely touch on consumer rights, such as financial planning, everyday services, and rights-protection choices, these providers should use appropriate methods to flag the limitations of AI-generated content and prompt users to double-check critical details. At the same time, they should continuously refine model precision, bolster content review, risk alerts, and error-correction mechanisms, and raise the standard of responses to cut down on misinformation and misleading content, thereby reducing the likelihood of consumers suffering losses from over-reliance on AI output.

The CCA also urges businesses that employ AI customer service tools to tighten their review and management processes. These tools must deliver information that stays consistent with the actual goods and services, pricing, activity rules, contractual terms, and after-sales policies, and businesses should offer smooth, effective human customer service and dispute resolution channels. While leveraging AI to boost operational efficiency, companies are expected to establish corresponding accountability frameworks. When AI customer service provides significant answers to consumers regarding pricing, costs, or after-sales matters, businesses cannot simply disclaim responsibility on the grounds of “algorithmic generation,” nor can they rely on such disclaimers to improperly weaken or avoid their legal obligations.

The CCA calls on relevant authorities to further refine the rules governing AI’s application in the consumer sector, clarifying the responsibilities and governance requirements of service providers and businesses across different service types and usage scenarios. This would foster an AI-driven consumer environment that is clearly defined, orderly, and accountable. Additionally, the CCA reminds consumers to take a rational view of general-purpose generative AI as a tool for reference and assistance only. When it comes to critical consumer information like pricing, contract terms, after-sales support, or legal liabilities, decisions should not rest solely on AI-generated content. Instead, consumers are encouraged to verify such details through official business channels, contract documents, regulatory bodies, or other trusted sources.