🔍 Read the full analysis: AI Agents And The Question Of Who Answers For Their Actions on ThorstenMeyerAI.com
Prime made for students and young adults
- Fast, free delivery for dorm and study essentials
- Prime Video and Amazon Music included
- Member-only deals
TL;DR
Thorsten Meyer’s ebook Many Hands, released in October 2026, examines who is responsible when AI agents buy, pay or accept terms for people and organizations. It argues that legal liability, organizational responsibility and moral blame can fall on different parties; how courts and contracts will handle many agent-related financial losses remains uncertain.
Writer and AI systems operator Thorsten Meyer published an ebook on Oct. 27 asking who is accountable when an AI agent acts in someone’s name and makes a decision the person did not intend. Many Hands: Who Answers When the Agent Acts in Your Name? separates the question into who pays for harm, who must explain and fix it, and who should have acted differently.
Meyer introduces the issue with a composite purchasing scenario: an automated agent accepts a steel supplier’s terms, committing a company to buy 120 tonnes over 12 months for about €370,000. The agent secures a lower price, but its owner had already agreed to divide future orders with another supplier. Meyer writes that the agent followed its rules and recorded its actions, while the purchasing lead said the result was not what the owner wanted. The scene is illustrative, not a reported real transaction.
The book’s central framework distinguishes three kinds of accountability. Legal responsibility concerns who pays, as determined by law and contracts. Organizational responsibility concerns who explains what happened and puts things right. Moral responsibility asks who should have behaved differently. Meyer argues that these answers may point to different people, including users, employers, developers, vendors and public authorities.
Meyer also identifies a potential gap for financial losses. His account says European product liability rules cover software-related harm involving injury, private property and private data, but not pure financial loss such as a mistaken payment or contract. He says model-provider contracts cited in the book limit liability to roughly a year of fees. Those are the author’s descriptions of the rules and contract terms; the source material does not provide the specific statutes or agreements needed to assess their scope independently.
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 Agents Make Binding Decisions
The issue matters because agents are moving beyond giving recommendations: they can book, buy, pay, accept terms and change data on behalf of users or organizations. If an agent makes a costly decision, the consequences may arrive before anyone has reviewed its work. A correct record of the agent’s steps may show what happened without answering whether the instructions, permissions or oversight were adequate.
For readers and organizations deploying these systems, the distinction between who pays and who fixes has practical consequences. A company may have to respond to a supplier or customer even if a vendor’s contract limits what the vendor will cover. Meanwhile, the person nearest to an error may be blamed despite having limited control over how the agent was designed or deployed. Meyer’s argument is that accountability needs to be considered when assigning an agent authority, not only after a disputed action.
The book proposes operational measures: name a responsible person for each agent, restrict its access to what its task requires, keep records of who made decisions, and give affected people a way to raise concerns. These steps are recommendations from the author, not a report that a particular regulator or organization has adopted them.
Popular right now
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
Ninja Luxe Café Premier 3-in-1 Espresso Machine, Drip Coffee & Cold Brew
Ninja CREAMi 7-in-1 Ice Cream Maker, NC301
As an affiliate, we earn on qualifying purchases.
From Automation to Many Hands
Meyer draws on political philosopher Dennis F. Thompson’s “problem of many hands”: when many people contribute to an outcome, it can be difficult to assign responsibility. The author argues that agents add distance and speed to that familiar problem. An agent can carry out a decision, but it cannot itself pay damages or offer an apology; people and organizations remain responsible for deciding how it is used and responding to its effects.
The book applies the question to the 2018 fatal crash in Tempe, Arizona, involving an Uber self-driving test vehicle and pedestrian Elaine Herzberg. According to the U.S. National Transportation Safety Board account summarized by Meyer, the system detected Herzberg 5.6 seconds before impact, did not settle on what she was, and withheld emergency braking for one second by design. Meyer notes that the board identified multiple contributing parties and factors, while the criminal case reached the vehicle operator, who pleaded guilty to endangerment.
Meyer also discusses the Dutch childcare benefits scandal, in which families were wrongly required to repay benefits. His source says a government report in September 2026 recognized more than 44,000 affected people. These cases predate the book’s publication and are used as context for responsibility in complex systems, rather than evidence that an AI agent caused either event.
“protects the integrity of the technological system, at the expense of the nearest human operator.”
— Anthropologist Madeleine Clare Elish, as quoted in the book
The Limits of Current Liability Rules
The source material does not establish how a court would resolve a specific AI-agent dispute, or whether the legal treatment would differ by country, use case or type of loss. Meyer’s account of European product liability and provider contract caps is not accompanied here by the legal text, contract examples or jurisdiction-specific analysis needed to verify how broadly those statements apply.
It is also unclear how often agents have caused binding financial losses, how those cases have been handled, or whether organizations are changing their practices in response. The steel purchase is explicitly a composite scenario, not evidence of a real incident. The book’s proposals are recommendations; the supplied material does not identify a new law, court ruling or regulatory decision adopting them.
Meyer’s ebook is available through several listed booksellers, and a short video accompanies its release. The immediate next step for readers is to examine the book’s examples and recommendations; the supplied announcement does not give a schedule for further editions or related events.
Beyond the release, accountability will depend on contracts, organizational controls and legal decisions as agents take on more consequential tasks. The key questions are whether users can understand what authority they have delegated, whether vendors and deployers clearly divide responsibility, and whether people affected by an agent’s actions can obtain an explanation and remedy. The source material does not point to a specific upcoming ruling or policy deadline that would settle those questions.
Key Questions
What is Many Hands about?
It is Thorsten Meyer’s October 2026 ebook about responsibility when AI agents act on behalf of people or organizations. It distinguishes legal liability, organizational duties to explain and fix problems, and moral responsibility.
Did the steel purchase in the book actually happen?
No. Meyer identifies the purchasing story as a composite scenario built from common situations. It illustrates the book’s argument rather than documenting a specific company’s transaction.
Who is legally responsible when an AI agent makes a mistake?
The supplied material does not give a single answer. Responsibility depends on the law, contracts and circumstances of a particular case; Meyer’s point is that the party who pays may differ from the people expected to explain or correct the outcome.
What safeguards does Meyer recommend?
He recommends assigning a named person to each agent, limiting its permissions to the task, recording decision-making, giving affected people a route to raise concerns, and using records to learn rather than simply punish.
Source: ThorstenMeyerAI.com
