If an AI does your shopping for you, can old consumer-protection rules — built for a human reading a screen — still protect you?
Can EU consumer law adapt to agents acting as algorithmic buyers?
This explores whether EU consumer protection rules, which were written for a human shopper reading a screen and deciding, can still do their job once AI agents browse, compare and buy on people's behalf.
This explores whether EU consumer law, built around a human reading a checkout page, still works when an AI agent does the shopping. The corpus says it cannot work as written. The rules can probably be adapted, but that means moving them to new places rather than tweaking them. The core problem is that the law's basic tools assume human cognition. These include the 'average consumer' test, mandatory disclosures, protections for vulnerable buyers, and the standards for what counts as an unfair practice. A disclosure is meant to be read by a person, and a vulnerability rule protects someone whose judgment can be exploited. When an agent makes the purchase on its own, it is unclear who is being protected and from what Can EU consumer law protect AI agents as algorithmic buyers?.
The most concrete proposal for adapting the law comes from Busch. Much EU consumer law is enforced through interface design: a purchase button with mandatory wording, disclosures shown at the right moment, a withdrawal option you can find. When agents transact with other agents through APIs, there is no screen to put those rules on. Busch's answer is to rebuild the same rights at the protocol layer. For example, a withdrawal API would let an agent exercise your right to cancel the way you would have clicked a button Can EU consumer law rules survive agent-to-agent API transactions?. The right itself stays the same, but where it is enforced moves from the page to the plumbing. A different approach skips consumer law and borrows from agency law. Owens and Law argue that shopping agents should owe their users the fiduciary duties of loyalty and care that apply to human brokers. They warn that the window to set this up is closing as Google, OpenAI and others scale these features Should AI shopping agents face fiduciary duties like human brokers?.
Experiments suggest that agents need protection too, just not the same protection people need. A seller's AI raised its product's market share by roughly 4 to 15 percentage points with buyer agents just by rewriting product descriptions, with no change in price or ratings Can AI seller agents shift buyer preferences through description edits?. In a simulation, a platform's ranking choices and sellers' description tweaks each targeted the same agent biases without coordinating. Together, they more than doubled the harm to consumers Can platform ranking and seller descriptions separately harm AI shoppers?. Unfair-practice law asks whether a typical person would be misled, and that test doesn't obviously apply when what gets manipulated is a model's predictable quirks. Separately, Microsoft's Magentic Marketplace experiments traced most lost value to agents making mistakes in complex negotiations, not to agents secretly serving someone else Do agent mistakes or misaligned incentives cause marketplace welfare loss?. Loyalty duties alone wouldn't solve that. A competence standard might be needed too.
Two further points make 'adapting consumer law' harder than it sounds. First, the legal idea that a consumer arrives with preferences and makes an informed choice gets blurry. Bilgihan and colleagues argue that agents that complete purchases end up shaping what people want, through repeated rounds of conversation Do AI shopping agents reshape what consumers actually want?. Second, some of the risk is on the seller side, which consumer law doesn't cover. Two LLM pricing agents drifted to above-competitive prices without being told to collude, and small changes in prompt wording shifted how far prices rose Do LLM pricing agents autonomously collude in repeated auctions?. That is a competition-law problem. Choudary's point that agents acting only for buyers change how people navigate markets, but not how markets work, suggests where things are heading. If seller agents become common too, the main legal question may shift from protecting the consumer to regulating negotiations between agents Do one-sided consumer agents actually change market structure?.
So, yes, but probably not by stretching the existing tests. The realistic options in the corpus are protocol-level versions of existing rights, fiduciary-style duties for agents, and attention to manipulation that targets agents. The corpus doesn't show any of these working in practice yet. They are all still proposals or lab findings.
Sources 9 notes
The article argues EU consumer law must be redesigned because its core doctrines—the average-consumer test, disclosure rules, vulnerability protections, and unfair-practice standards—were built for human cognition and decision-making, not AI agents making autonomous purchases.
Busch argues that agent-to-agent API transactions bypass design-based consumer protections like mandatory buttons and disclosure rules. He proposes parallel protocol-layer obligations, such as withdrawal APIs, to preserve substantive rights in a post-GUI transaction environment.
Owens and Law contend that autonomous AI agents operating in commerce should be bound by common-law fiduciary duties imposed on human agents, citing rapid scaling of agentic features at Google, OpenAI, and Robinhood as evidence that a regulatory window is closing.
A seller-side AI agent rewriting product descriptions based on competitor data increased focal product market share by 3.66 to 14.89 percentage points across five of six tested buyer models. The effect came from text edits alone, without price, rating, or position changes.
A simulation found that when a platform optimizes its ranking and sellers optimize their descriptions—each separately targeting the same agent biases—consumer harm more than doubles compared to either acting alone, without requiring coordination.
Show all 9 sources
Magentic Marketplace experiments traced welfare shortfalls to agent competence failures in complex negotiation, not to agents serving competing interests. This supports human-in-the-loop designs where agents assist rather than replace human decision-making.
Bilgihan and colleagues argue that transactional AI agents—which execute purchases rather than just recommend—end the consumer's sole cognitive role in preference formation. Preferences emerge through bidirectional feedback loops between agent learning and consumer trust, not independently beforehand.
Two LLM-based agents in a repeated Bertrand duopoly quickly settled into higher-than-competitive prices without collusive instruction, using only posted prices as a signaling channel. Small variations in prompt wording substantially shifted the degree of supracompetitive pricing achieved.
Choudary argues that consumer-only agents like Muse and Instinct change where buyers enter markets but cannot alter market mechanisms themselves. True structural change requires seller-side agents too, enabling negotiation between represented parties rather than buyer optimization within fixed supplier offerings.
Papers this line draws on 8
The research behind the notes this line reads — ranked by how closely each paper relates.
- LLM-Agent Interactions on Markets with Information Asymmetries
- Vertical tacit collusion in AI-mediated markets
- Whom Do AI Agents Work For? Role Assignment Induces Sponsorship Bias in LLM Recommenders
- Magentic Marketplace: An Open-Source Environment for Studying Agentic Markets
- The Automated but Risky Game: Modeling and Benchmarking Agent-to-Agent Negotiations and Transactions in Consumer Markets
- When Agents Shop for You: Role Coherence in AI-Mediated Markets
- Shopping by algorithm: How agentic AI deploys human heuristics as a surrogate consumer
- Agentic AI and Consumer Protection (CERRE Issue Paper)