🔍 Read the full analysis: When AI Acts On Your Behalf, Who Should Be Held Responsible? 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’s ebook Many Hands, published in October 2026, examines who should answer when AI agents make purchases, accept terms or otherwise act for people and organizations. Its central argument separates who pays, who must explain and fix a failure, and who bears moral responsibility; the legal and practical answers can differ.
Author and AI systems builder Thorsten Meyer published Many Hands in October 2026, an ebook about who should answer when an AI agent acts in a person’s or company’s name. The book argues that responsibility has three distinct parts—who pays, who explains and repairs the harm, and who should have acted differently—which may not fall on the same person.
Meyer frames the issue through a purchasing scenario involving Theresa, a composite character rather than a reported real case. An agent accepts a steel supplier’s terms, securing a lower price but committing the company to buy 120 tonnes over a year without a cancellation right. The commitment is valued at about €370,000. Theresa had previously agreed to divide the next year’s volume with another supplier, but the agent had not been informed. Meyer says the agent followed its rules and recorded its steps, even though its action conflicted with the owner’s plans.
The book’s central distinction is between legal liability, internal responsibility for explaining and fixing a failure, and moral blame. Meyer argues that a transaction can comply with an agent’s instructions while still producing an outcome its user did not want. He also says blame can gravitate toward the closest, most visible human, while designers and organizational decision-makers remain harder to identify.
Meyer contends that existing rules and contracts can leave gaps. He says the EU’s product liability framework covers software but does not cover pure financial loss, such as a mistaken payment or contract. He also says provider terms cited in the book can limit liability to roughly a year of fees. These are the author’s descriptions and arguments; the supplied material does not include the relevant contract language or a legal analysis independently establishing how those rules would apply to a particular agent-related dispute.
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
Where Agent Errors Leave the Loss
AI agents are increasingly designed not only to generate answers but to take actions: buying goods, booking services, paying bills, accepting terms and changing records. When an agent acts with authority delegated by a person or business, a mistake can create a binding commitment or financial loss before anyone notices. The practical question is not just whether the software behaved as programmed, but who had the ability and duty to prevent, review or correct the action.
Meyer’s framework matters because a single label such as “responsible” can hide different outcomes. A supplier or customer may seek payment from one party, an organization may need to identify who can resolve the problem, and a moral assessment may point to people who designed policies or delegated authority. Those questions can lead to different answers. For organizations using agents, the book’s focus on named responsibility, limited permissions and reviewable records offers a way to think about governance, though the source does not establish that these measures alone prevent disputes or determine liability.
Products Worth Considering
Sony WH-1000XM6 Wireless Noise-Canceling Over-Ear Headphones, Blue, Bundle
Amazon Kindle Paperwhite Signature Edition (12th Generation, 2024), 32 GB, Metallic Black
Bose TV Speaker (2nd Gen) Compact Soundbar with Bluetooth and HDMI-ARC, Black
As an affiliate, we earn on qualifying purchases.
From Automated Decisions to Delegated Action
Meyer connects the problem to political philosopher Dennis F. Thompson’s 1980 phrase, “the problem of many hands”: when many people contribute to an outcome, it can be difficult to determine who is accountable. In the author’s account, agents add distance, speed and another decision-making layer to that familiar organizational problem. They do not remove the people who build, deploy or authorize them, but they can make it harder to see how those choices shaped a specific action.
The book uses two earlier examples to illustrate different forms of diffuse accountability. In the 2018 Tempe, Arizona, crash involving a self-driving test vehicle, Meyer cites the US National Transportation Safety Board’s findings about detection and braking, as well as its identification of multiple contributing factors. The book also discusses the Dutch childcare benefits scandal, in which families were wrongly required to repay benefits. Meyer says the Dutch government recognized more than 44,000 affected people by September 2026. These cases concern different technologies and institutions; they provide context for the book’s argument, not direct evidence about liability for today’s AI agents.
“protects the integrity of the technological system, at the expense of the nearest human operator.”
— Madeleine Clare Elish, as quoted in Meyer’s book
How Agent Liability Will Be Decided
The source material does not identify a court ruling that settles who is liable when an AI agent accepts a contract or makes a mistaken payment. How liability would apply depends on the facts, applicable law, and contractual terms; the book’s account of EU product-liability rules and provider limits is not a substitute for case-specific legal advice. It is also unclear how frequently agent actions have produced comparable losses, since the source provides no industry-wide data.
The steel purchase is expressly a composite illustration, so it should not be treated as evidence that a specific company incurred a €370,000 loss. The source also does not provide independent confirmation of the book’s account of current vendor contract terms or describe the outcome of any lawsuit involving the scenario.
What Organizations Can Put in Place
Meyer recommends assigning a named person to each agent, limiting its permissions to the task, keeping records of who made and authorized decisions, and giving affected people a route to raise problems. He also argues that records should support learning rather than serve only as a basis for punishment. These are the author’s proposed safeguards, not requirements shown in the source to be binding law.
The immediate development is the ebook’s publication; the source does not announce a hearing, regulatory decision or scheduled policy change. The broader questions—how courts will allocate losses, how businesses will write contracts for agent-driven transactions, and how organizations will supervise delegated actions—remain open. Readers can expect those answers to depend on future disputes and rules, rather than on the book’s argument alone.
Key Questions
What is Many Hands about?
Many Hands is Thorsten Meyer’s October 2026 ebook about responsibility when AI agents act for people or organizations. It separates legal liability, organizational duties to explain and fix a failure, and moral responsibility.
Is the steel-purchasing story a real case?
No. Meyer describes Theresa and the purchasing scenario as a composite example built from common situations. The source does not identify it as a real company’s transaction.
Does the book establish who is legally liable for an agent’s mistake?
No. It raises legal questions and discusses potential gaps in rules and contracts, but the source does not report a ruling that resolves liability for AI-agent actions. The outcome of a dispute would depend on its facts and applicable law.
What safeguards does Meyer recommend?
He proposes assigning each agent a named responsible person, limiting permissions, keeping decision records and giving affected people a way to raise concerns. He also says records should be used to improve systems, not only to punish individuals.
Source: ThorstenMeyerAI.com
