Consumer Law for AI Agents

Introduction. On January 23, 2025, OpenAI, the company behind ChatGPT released a preview version of “Operator”, an AI agent that can browse the web and perform tasks for users.1 Operator can interact with websites and automate a variety of actions like buying goods online, booking train tickets and making restaurant reservations. Unlike AI chatbots, agentic AI systems such as Operator are not limited to image generation or answering questions. They can interact with digital environments and take actions on behalf of users. In particular, consumers could soon be able to delegate their purchasing decisions to AI agents. Sooner rather than later, AI agents will interact autonomously with apps or online shops and purchase goods and services in accordance with pre-defined preferences and payment options. Against this background, this Article argues that EU consumer law needs to prepare for the rise of AI agents such as OpenAI’s Operator and even more powerful AI tools that might become available in the near future. Consumer law, as it currently stands, is human-centric. It is based on the premise that decisions to buy goods or services are made by humans. In the future, however, a growing number of consumption decisions will be made by AI agents acting as “algorithmic consumers”2, “machine customers”3 or “Custobots”4 (a portmanteau of ‘customer’ and ‘robot’). Consumer policy makers need a plan to update the existing regulatory framework to enable consumers to capture the benefits of AI agents and to mitigate potential risks arising in the context of agentic commerce. This impending shift in consumer decision-making from humans to AI agents requires several adjustments to the regulatory design of consumer law. Currently, consumer law is geared towards enabling informed decision-making by human consumers. In the future, however, the specific strengths and weaknesses of non-human consumers must be taken into account. As a consequence, for example, the design of pre-contractual disclosures has to be reconsidered as AI agents can process larger amounts of information. Similarly, the rules on unfair commercial practices need to be recalibrated as the vulnerability of AI agents to misleading and aggressive commercial practices may differ from that of humans. On the one hand, AI agents will be less vulnerable to dark patterns and emotional advertising messages. On the other hand, new AIspecific vulnerabilities could arise such as prompt injections and other adversarial attacks that trick AI agents into a transaction. This article adds to the expanding literature on AI agents. Other scholars have analyzed legal issues regarding AI agents through the lens of agency law5 or they focus primarily on issues of liability, antitrust or macro-level effects of AI agents.6 This article chooses a different perspective. Building on the work of the European Law Institute’s project on Algorithmic Contracts,7 this Article explores whether EU consumer law, as it currently stands, is ready for the rise of AI agents. In doing so, the Article seeks to identify implications for consumer policy, and motivate further empirical research. The remainder of the Article proceeds as follows. Part B provides a brief overview of the recent development from generative AI to agentic AI and how AI agents will change the functioning of consumer markets. Part C turns to the question what the rise of AI agents means for consumer law. In particular, it will be considered how the emergence of the Custobot Economy will challenge the premises of consumer law. Part D seeks to identify core elements that could contribute to a consumer law that works for both humans and machines. Part E concludes.

I. From Chatbots to Custobots Since the public release of ChatGPT in November 2022, generative AI has taken the world by storm. ChatGPT, which reached more than 100 million monthly users in just two months, is widely seen as the fastest-growing consumer internet app of all time.8 More recently, generative AI tools have become relevant for e-commerce. Consumers can ask tools like OpenAI’s ChatGPT, Microsoft’s Copilot, or Google’s Gemini for advice, product recommendation or suggestions where to buy goods and services. In November 2024, the AI-driven search engine Perplexity introduced a new feature that allows Pro subscribers to purchase a product without leaving its AI search engine. Users can click a “Buy with Pro” button that will automatically order the product on the basis of saved shipping and billing information.9 AI agents like OpenAI’s Operator are taking the shopping integration one step further. In the words of Jonathan Zittrain, the advent of AI agents marks “a crossing of the blood-brain barrier between digital and analog, bits and atoms”.10 Unlike AI-driven chatbots that are designed to assist users in their transactional decisions, AI agents can act on behalf of users.

Method. II. The Rise of Agentic Commerce The rise of AI agents represents a rare moment of technological disruption—one with the potential to fundamentally reshape how consumers engage with online marketplaces. This shift towards agent-driven commerce may prompt a sweeping reconfiguration of the current ecommerce landscape. In turn, it raises key conceptual questions about the nature and mechanics of online shopping, suggesting that some foundational assumptions about how digital markets function may need to be reconsidered. To be sure, e-commerce has evolved continually since Amazon and eBay first launched their online marketplaces in 1995. Over time, user interface design has undergone several shifts, moving from early websites to mobile commerce, social commerce, and more recently, live shopping on platforms like TikTok. Yet throughout these waves of technological change and experimentation with business models, one element has remained constant: user interfaces were always designed to persuade human consumers to make purchasing decisions. This core feature of e-commerce is likely to change with the advent of AI agents and the rise of agentic commerce (“a-commerce”).19 As human consumers increasingly delegate purchasing decisions to autonomous AI systems, the focus of online marketplaces will move from persuading humans to influencing AI agents. This transition is set to introduce several fundamental changes to the structure of e-commerce—each carrying significant implications from a consumer policy perspective.

1. New Customer Journeys First of all, AI agents will change the way customers search for products. Today, the customer journey often begins with a search engine like Google or one of the large online marketplaces, such as Amazon, which has itself become a kind of product search engine. In the near future, AI agents could serve as new access points for commerce.20 Deutsche Telekom recently presented the concept of an “app-less” AI phone which gives an idea of how this new generation of user interfaces might look like.21 On the so-called “T-phone”, users interact with service providers through an AI agent rather than individual apps. Users present their questions to the AI agent, which autonomously selects the underlying apps or service providers needed to fulfill the query. Rather than starting their customer journey at a specific website, app or search engine consumers will instruct their AI agent to find a product that matches their preferences and order it on their behalf. In this sense, AI agents are not merely a new “technology”.22 They will change how consumers interact with websites and online shops. From a competition perspective, this change has the potential to disrupt the “gatekeepers” that are controlling today’s access points to online retail.23 The developments outlined here could have far-reaching consequences for the design of online shops. The internet, website and apps largely have been built for use by humans and optimized to improve user experience for human customers. However, the shift towards agent-to-website interaction or agent-to-agent interaction could require major changes in the design of user interfaces. In particular, the content offered via websites and apps will need to focus more on information that AI agents can use to perform their tasks. In this sense, customer-centricity will require to understand how AI agents interact with websites and apps. It remains to be seen what specific impact this will have on the design of online shops. One possible scenario is that that websites and apps will become the target for both human and agentic traffic. It is also conceivable that agent-to-agent exchanges could be driven by API connections which will be more scalable and can be better adapted to the information needs of AI agents. As a consequence, a new category of online shops for Custobots could emerge that resemble “dark supermarkets” or “ghost kitchens” specifically designed to cater for delivery services.

Discussion. C. How AI Agents Challenge the Premises of Consumer Law Building on the analysis of the upcoming changes of the e-commerce landscape, this Part turns to the question why the rise of AI agents challenges the basic premises of consumer law and how consumer law rules must adapt to the new reality of AI driven markets. As outlined above, the advent of AI agents has the potential to fundamentally reshape how consumers make purchasing decisions or other transactional decisions (e.g. managing subscriptions). If built with the right legal and technical safeguards, AI agents could become a tool for consumer empowerment shifting the balance of power between businesses and consumers.35 In so doing, AI agents could save consumers considerable time and money.36 More specifically, AI agents could help consumers make better informed choices, and less biased purchasing decisions. Does this mean that AI agents will finally embody the theoretical model of the rational homo economicus? Not necessarily. The answer is more complicated, as will be explained in the following sections. In doing so, the focus will be on four topics: the average consumer concept, information duties, consumer choice and the concept of digital vulnerability.

I. Beyond human-centricity: Towards an Average Consumer Test for Custobots? For assessing the profound impact of AI agents on EU consumer law, it is necessary to consider first a fundamental premise of consumer law that seems so self-evident that it is usually not even stated. EU consumer law in its current form is human-centric, i.e. it is designed with the needs and weaknesses of human consumers in mind. Put differently, the design of consumer law rules is based on the premise that purchasing decisions are made by humans. Therefore, EU consumer law aims to protect human consumers and ensure their autonomy. But this premise no longer holds in the Custobot Economy, which will be more complex and diverse. In the brave new world of AI-driven retail old-school human consumers will coexist with “augmented consumers” (who are assisted by AI tools such as ChatGPT) and AI agents acting as “algorithmic consumers”. The shift towards agentic AI systems could have a particular impact on the concept of the “average consumer”. This consumer image or Verbraucherleitbild serves as a normative standard for the application of consumer law rules.37 For example, an assessment of whether a commercial practice is unfair involves an examination of whether it materially distorts or is likely to materially distort the economic behavior of an average consumer.38 To answer this question, the Court of Justice of the European Union (CJEU), takes as a benchmark the average consumer who is “reasonably well-informed, reasonably observant and circumspect”.39 This average consumer concept, which was initially developed by the case law of the CJEU, has since been codified by the European legislator in the secondary legislation.40 The average consumer concept has long been criticized by behavioralists as an unrealistic assumption.41 Indeed, having a realistic understanding of consumer behavior is critical for getting regulation right.

Conclusion. Scholars have been arguing already for some time that the one-size-fits all model of consumer law no longer corresponds to the complexities of modern consumer markets and that a greater differentiation between different types of consumers is required.97 The rise of AI agents brings a new twist to this argument. With the rise of agentic commerce, the complexity of the consumer market is increasing even further. In the new Custobot Economy, human consumers will coexist with AI-driven “algorithmic consumers”. This will affect almost all aspects of consumer law, from the average consumer concept to the information paradigm and the concept of digital vulnerability. Adapting consumer law rules to the coming wave of Custobots will require a nuanced understanding of the strengths and weaknesses of AI agents. Policymakers should not wait for the next “iPhone moment”98 to address these challenges; they should start now creating a consumer law that works for both humans and machines.