🔍 Read the full analysis: My New Book Follows The Question Of AI Agent Responsibility 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 says his ebook Many Hands: Who Answers When the Agent Acts in Your Name? was published in October 2026. It examines how legal liability, organizational responsibility and moral blame can fall on different people when AI agents act for them. The source describes the book’s arguments and examples, but does not establish how courts will resolve the issues for AI agents.
Thorsten Meyer says his ebook Many Hands: Who Answers When the Agent Acts in Your Name? was published in October 2026, focusing on who pays, explains and bears moral responsibility when an AI agent acts for a person or organization. The book uses a fictional purchasing scenario and cases involving automated decision-making to argue that those responsibilities can fall on different parties.
The opening scenario follows Theresa, a business owner, whose purchasing agent accepts a steel supplier’s terms overnight. The agent secures a 7% lower price per tonne but commits the company to buying 120 tonnes over 12 months, without a cancellation right. Meyer puts the commitment at about €370,000. The day before, Theresa had agreed to divide the next year’s volume with another supplier, but the agent was not given that information. Meyer identifies Theresa as a composite character, not a real person in a reported transaction.
The book frames the resulting question—who bought the steel?—as three questions: who pays under law and contracts, who inside an organization must explain and address the outcome, and who should have acted differently. Meyer argues that these answers may diverge when an agent follows its programmed rules but acts against a user’s intent. The source says the agent in the scenario followed its rules and recorded its steps; it does not describe a real dispute or legal ruling.
Meyer also discusses the 2018 fatal crash involving an Uber self-driving test vehicle in Tempe, Arizona, and the Dutch childcare benefits scandal. These are presented as examples of responsibility becoming concentrated on a nearby person or dispersed across institutions. The book is an ebook, with the first edition dated October 2026; the author lists several retailers, including Apple Books, Kobo and Barnes & Noble.
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
Responsibility Beyond the Agent
The subject matters as agents take on actions with financial and operational consequences, including purchases, payments, bookings and changes to data. When an agent acts for someone, a record showing that it followed its rules may not settle whether the action was authorized, who bears the loss, or who must put things right.
Meyer’s distinction among legal, organizational and moral responsibility offers readers a way to separate those questions rather than treating “accountability” as a single answer. His proposed safeguards include a named person for each agent, permissions limited to the task, records showing who made decisions, and a route for affected people to raise concerns. These are recommendations presented by the author, not requirements shown to have been adopted across the industry.
The issue also has practical consequences for businesses evaluating agent systems. A favorable price or efficient action may still create an unwanted commitment if an agent lacks relevant information or authority limits. The book’s scenario illustrates why oversight and the ability to refuse or reverse consequential actions can matter alongside a system’s technical performance.
Popular right now
Apple AirPods Pro 3 Wireless Earbuds with Active Noise Cancellation
Bose QuietComfort Wireless Bluetooth Headphones with Active Noise Cancellation, Mic, USB‑C Charging, Deep Bass, 24‑Hour Battery – White Smok
Ninja CREAMi 7-in-1 Ice Cream Maker, NC301
As an affiliate, we earn on qualifying purchases.
From Agent Tools to Agent Actions
Meyer says his earlier writing addressed agent systems, including credentials and shopping agents. He presents this book as a further question: who answers when an agent acts in someone’s name? That question becomes more pressing as systems move beyond generating responses and begin interacting with services or making decisions that can affect contracts and money.
The book draws on the idea of the “problem of many hands,” a term associated with political philosopher Dennis F. Thompson’s 1980 writing about how responsibility becomes difficult to assign when many people contribute to an outcome. Meyer argues that agents add distance and speed to this problem. The source attributes the phrase to Thompson but does not provide publication details beyond the year.
For the self-driving-car example, Meyer cites the US National Transportation Safety Board’s account of the 2018 Tempe crash, saying the system detected Elaine Herzberg 5.6 seconds before impact but did not settle on what she was and withheld emergency braking for one second by design. He also cites the Dutch childcare benefits scandal and a September 2026 government report that he says recognized more than 44,000 people as affected. These examples provide background for the book’s argument; they are not cases about AI agents making purchases.
“One question has three answers.”
— Thorsten Meyer, author of Many Hands
How Agent Liability Will Be Settled
The source announces and describes Meyer’s book; it does not report a new court decision, regulatory action or independently assessed finding about AI agent liability. It remains unclear how courts will apply existing laws when an agent accepts terms or makes a purchase without the user’s intended authorization. The source also does not identify a real company, supplier or transaction behind the steel scenario, which Meyer calls composite.
Meyer argues that European product-liability rules cover software but leave out pure financial losses such as a bad contract or incorrect payment. The source does not provide the legal text or analysis needed to assess that account, and the scope of liability may depend on the specific law, contract and facts. It also says model-provider contracts may cap liability at about a year of fees, but supplies no named providers or contract examples in the material provided. The claims should be read as the author’s account of legal and contractual risks, not a ruling that applies to every product or agreement.
The source does not give independent evidence about the book’s sales, reception or the degree to which its suggested safeguards are being used. It also ends partway through its account of an Amsterdam Court of Appeal case from April 2023, so the case’s details and relevance cannot be established from the supplied material.
Where the Accountability Debate Goes
Meyer’s source points readers to the ebook, its listed retailers and a short video introducing the book’s argument. It does not announce a launch event, planned follow-up research, or a timetable for further editions. No response from regulators, AI companies or legal experts is included.
The next developments that could clarify the broader issue are specific legal decisions, regulatory guidance and contract terms addressing when an agent is authorized to act and who bears losses when it exceeds a user’s intent. Until those emerge, readers and organizations will need to distinguish the book’s proposals from established legal duties. Meyer’s stated practical focus is on assigning a responsible person, setting permissions for the task and keeping decision records that can support review and correction.
Key Questions
What is Many Hands about?
Many Hands examines who is responsible when an AI agent acts for a person or organization. Meyer separates the questions of who pays for a loss, who must explain and fix the outcome, and who should have acted differently.
Is the steel purchase in the book a real incident?
No. Meyer describes Theresa as a composite character and the steel purchase as an opening scenario, not a reported real transaction.
When was the ebook published?
The source lists the first edition as October 2026 and says the ebook is available through retailers including Apple Books, Kobo and Barnes & Noble.
Does the book establish who is legally liable for an agent’s actions?
No. The source presents questions and arguments about liability, contracts and oversight; it does not report a legal ruling that settles responsibility for AI agents generally. The answer in any dispute may depend on the relevant law, agreement and facts.
What safeguards does Meyer propose?
He recommends naming a person responsible for each agent, limiting its permissions to the assigned task, recording who made decisions, and providing a way for affected people to raise concerns. These are the author’s proposals, not evidence of a universal standard.
Source: ThorstenMeyerAI.com
