🔍 Read the full analysis: What Does It Mean For AI To Act In Your Name? on ThorstenMeyerAI.com
Get the latest gadgets delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Thorsten Meyer published Many Hands: Who Answers When the Agent Acts in Your Name? as an ebook in October 2026. The book argues that agent accountability has three separate parts—who pays, who must explain and fix an outcome, and who should be held morally responsible—and says current law and contracts may leave financial losses difficult to recover.
Thorsten Meyer released Many Hands: Who Answers When the Agent Acts in Your Name? as an ebook in October 2026, focusing on how responsibility is divided when an AI agent takes action for a person or business. The book distinguishes who pays for a loss, who must explain and repair the outcome, and who bears moral responsibility—questions that can have different answers when systems act under delegated authority.
Meyer opens with a fictional purchasing scenario: an agent accepts a steel supplier’s terms overnight, securing a 7% lower price per tonne but committing a company to buy 120 tonnes over 12 months, with no cancellation right. The commitment is valued at about €370,000. The owner, Theresa, had agreed the previous day to split next year’s volume with another supplier, but the agent did not have that information. Meyer identifies Theresa as a composite character, not a real person or a documented transaction.
The scenario captures the book’s central distinction: an agent can follow its stated rules and keep a record of its actions while still producing an outcome its principal did not want. Meyer frames accountability in three ways: legal responsibility for who pays, organizational responsibility for who explains and fixes a failure, and moral responsibility for who should have acted differently. Those responsibilities may fall on different people or organizations.
The author argues that practical safeguards can make delegation more accountable. His proposed measures include naming a person responsible for each agent, limiting its permissions to the task, recording who made decisions, providing a way for affected people to raise concerns, and using logs to improve systems rather than only to punish staff. These are recommendations in the book; the source material does not establish that they are requirements under a particular law.
Many Hands · the core idea
One question, three answers
"Am I liable if the agent does it in my name?" sounds like one question. The book splits it into three, because agent systems pull the answers apart.
Decided by courts, contracts and insurers, on their own timetable.
In Theresa's caseHer company will almost certainly pay for the steel, whatever it later recovers from others.
Decided inside the organization: who must say what happened and what will change.
In Theresa's caseTheresa explains it to the second supplier, her staff and perhaps her bank.
Who owes the people affected an acknowledgment. Often answered last, if at all.
In Theresa's caseStill open when she reaches the motorway.
The hands behind one overnight order
Introduction · a composite scene from the book. None of them saw the acceptance before it went out.
- Model makertrained the language model
- Vendorbuilt the agent on top of it
- Consultantwrote its rules with her staff
- Supplier's portalaccepted the declaration at 2:14 a.m.
- Her companyswitched it on; Theresa signed the rules
“The moral crumple zone protects the integrity of the technological system, at the expense of the nearest human operator.”Madeleine Clare Elish (2019), quoted in Chapter 1
Source: Thorsten Meyer, Many Hands: Who Answers When the Agent Acts in Your Name? (ebook, October 2026; law as of 16 October 2026), Introduction and Chapter 1. Theresa is a composite built from common situations, labeled as such in the book.
Powered by Thorsten Meyer AI
When Delegated Actions Create Losses
The subject matters as AI tools move beyond answering questions to booking, buying, paying, accepting terms and changing data. When an action creates a contractual or financial consequence, a technically correct execution may still conflict with a person’s intention. The question is not only whether the system worked as designed, but who had authority to delegate the task, what information the agent had, and whether anyone could stop the action.
Meyer’s distinction also points to a practical gap between being liable and being accountable inside an organization. A court or contract may determine who bears a cost, while employees and customers still need a clear explanation and a route to correct the outcome. If responsibility is spread across developers, vendors, managers and operators, the nearest available person may end up answering for a decision they did not fully control.
The book argues that financial loss deserves particular attention. According to Meyer, Europe’s revised product liability rules cover software and certain harms involving injury, private property and private data, but do not cover pure financial loss, such as a bad contract or incorrect payment. That is the author’s account of the rules, not a legal opinion in this article; readers facing a dispute would need advice based on the relevant jurisdiction and facts.
Products Worth Considering
Samsung 27-inch Odyssey OLED G6 QD-OLED QHD Gaming Monitor
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
LG 34GX90SA-W 34-inch Ultragear WQHD OLED Curved Gaming Monitor
As an affiliate, we earn on qualifying purchases.
The Many Hands Behind an Agent
The book’s title draws on political philosopher Dennis F. Thompson’s 1980 description of the “problem of many hands”: when many people contribute to an outcome, it can be hard to identify who is responsible. Meyer argues that AI agents add distance and speed to an existing problem, while introducing systems that can take actions but cannot themselves pay compensation or apologize.
To show how accountability can become concentrated on the person nearest an outcome, Meyer discusses the 2018 fatal self-driving test-car crash in Tempe, Arizona. He cites the US National Transportation Safety Board’s findings that the system detected pedestrian Elaine Herzberg 5.6 seconds before impact, did not settle on what she was, and was designed to delay emergency braking for one second. The board identified multiple factors, including the vehicle operator, company oversight and safety practices, the state and the pedestrian. Meyer also notes that the operator pleaded guilty to endangerment, and uses the case to ask why a system’s hardest task was left to its human monitor.
Another example is the Dutch childcare benefits scandal, in which tens of thousands of families were wrongly made to repay benefits. Meyer writes that responsibility was distributed across people and institutions involved in the risk model, fraud processes, courts and government. He cites a September 2026 government report recognizing more than 44,000 affected people, and a January 2021 statement by the prime minister that responsibility had to be borne jointly if the whole system had failed.
“If the whole system has failed, responsibility can only be borne jointly.”
— The Dutch prime minister, quoted in Meyer’s account of the childcare benefits scandal
Legal Exposure Depends on the Facts
The source material describes the book’s arguments but does not provide a complete account of how existing laws apply to particular AI-agent transactions. Legal responsibility will depend on the jurisdiction, contract, delegated authority and circumstances; the book’s general discussion should not be read as a ruling on any specific dispute. Its claim about the scope of European product liability rules is presented here as Meyer’s analysis.
The opening steel-purchase scene is explicitly a composite example, so it does not establish that a real company made such a purchase or suffered that loss. The supplied material also does not identify a particular court decision on agent-made commercial contracts or describe a final legislative response to the issues the book raises.
It remains unclear how widespread costly agent errors are, how often vendor contracts limit recovery, and how organizations will assign oversight as systems gain authority to act. The source supplies no market-wide data or independent evaluation of the safeguards Meyer recommends.
Readers Can Examine the Book’s Case
Many Hands is available as an ebook, according to Meyer’s October 27, 2026, announcement, through retailers including Apple Books, Barnes & Noble, Kobo, Thalia, Bookshop.org, Smashwords, Fable and Vivlio. The author also points readers to a five-minute video about the book; its narration uses an AI voice clone of Meyer, according to the source.
The announcement does not specify a release schedule for a print edition, new regulation or other formal policy response. For organizations using agents now, the immediate questions raised by the book are operational: what authority an agent has, who is responsible for it, what information it can access, and how a person affected by an automated action can seek an explanation or correction.
Key Questions
What does it mean for an AI agent to act in your name?
It means the agent is authorized or configured to take actions for a person or organization, such as making a purchase, accepting terms or changing records. The source does not set out a single legal definition; the scope of authority depends on the arrangement and applicable law.
What is the main argument of Many Hands?
Meyer separates accountability into who pays, who must explain and fix the result, and who should have acted differently. The book argues that those responsibilities can be spread across people, companies and systems.
Is the €370,000 steel purchase a real incident?
No. Meyer says Theresa and the purchasing scenario are a composite example built from common situations. The source does not identify a real company or transaction behind it.
Does European product liability law cover losses caused by AI agents?
The source says Meyer argues that Europe’s revised product liability rules cover software and certain harms but not pure financial loss, such as a bad contract or incorrect payment. How a particular loss is treated depends on the law and facts that apply; the source does not provide a case-specific legal determination.
What safeguards does Meyer recommend?
He proposes assigning a named person to each agent, limiting permissions to the task, keeping records of decisions, giving affected people a way to raise concerns, and using logs to learn from failures. These are recommendations described in the book, not stated legal mandates.
Source: ThorstenMeyerAI.com
