🔍 Read the full analysis: My Book Looks At Responsibility For AI Acting On Our Behalf on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Thorsten Meyer’s 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 accountability can split when AI agents act for people or businesses, and argues that responsibility should be supported by clear human oversight and records.
Writer and AI systems developer Thorsten Meyer published an ebook on October 27 asking who should pay for, explain or take moral responsibility for decisions made by AI agents acting on someone else’s behalf. Many Hands: Who Answers When the Agent Acts in Your Name? focuses on agents that can buy, book, pay, accept terms and change data, actions that can create consequences beyond a conversation.
Meyer opens with a fictional composite: a purchasing agent accepts a steel supplier’s terms overnight, securing a 7% lower price while committing a company to buy 120 tonnes over 12 months, with no cancellation right. The commitment is worth about €370,000. The owner had agreed the day before to divide next year’s volume with another supplier, but the agent was not given that information. Meyer says the scene is constructed from common situations, not a documented individual case.
The book’s central distinction is between who pays, a legal question; who explains and fixes a failure within an organization; and who should have acted differently, a moral question. Meyer argues that these answers can fall on different people. A person nearest to an automated system’s action may be easiest to identify, while the people who designed, deployed or governed the system may be less visible.
Meyer proposes practical accountability measures: designate a person responsible for each agent, limit the agent’s permissions to the task, preserve records of decisions, give affected people a way to raise concerns, and use incident records to improve systems rather than simply punish staff. These are the author’s recommendations; the source material does not present them as legal requirements or as independently evaluated standards.
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 Final Click
The issue matters because AI agents can make commitments in a person’s or company’s name, while authority over their actions is spread across users, employers, software vendors and system designers. When an agent accepts a contract or sends a payment, the resulting loss may be immediate, but it may not be clear who had the information or authority needed to prevent it.
Meyer’s three-part framework helps distinguish legal liability from operational responsibility and moral blame. That distinction is relevant to businesses deciding how much autonomy to give agents, and to people affected by an automated decision who need to know where to seek an explanation or remedy. The book also raises a practical question for organizations: whether a human approval step is meaningful if the person asked to approve an action lacks the time, expertise, information or authority to reject it.
Products Worth Considering
Bose QuietComfort Wireless Noise-Cancelling Over-Ear Headphones, Black
Amazon Kindle Paperwhite (12th Generation, 2024), 16 GB, 7-inch Display, Jade
Amazon Kindle Paperwhite Signature Edition (12th Generation, 2024), 32 GB, Metallic Black
As an affiliate, we earn on qualifying purchases.
From Automation to Delegated Action
Meyer says his earlier work addressed risks involving agent credentials and online shopping systems. This book shifts attention from the mechanics of access to the responsibility that follows when an agent uses that access. Its title draws on political philosopher Dennis F. Thompson’s 1980 account of the “problem of many hands”: when many people contribute to an outcome, assigning responsibility can be difficult. Meyer argues that agents add distance and speed to that existing problem.
The book also refers to the 2018 fatal crash involving an Uber self-driving test vehicle in Tempe, Arizona. According to the US National Transportation Safety Board, 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 by one second. The board identified several contributing factors, including the vehicle operator, the company’s risk assessment and safety oversight, and the pedestrian. Meyer uses the case to examine how consequences can be concentrated on the person closest to an automated system even when decisions were shaped by a wider organization.
Another example is the Dutch childcare benefits scandal, in which tens of thousands of families were wrongly required to repay benefits. Meyer describes it as a case where responsibility was distributed among institutions and individuals. The source says the Dutch government’s September 2026 report recognized more than 44,000 affected people; that figure is attributed to the source material and is not independently verified here.
“If the whole system has failed, responsibility can only be borne jointly.”
— Dutch prime minister, quoted in Meyer’s account of the childcare benefits scandal
Where the Liability Gap Remains
The ebook’s discussion of law is the author’s analysis, not a ruling on any particular AI-agent dispute. Meyer says European product-liability rules cover software and address injury, private-property damage and private data, but do not cover pure financial loss, such as a mistaken payment or an unfavorable contract. The source material does not provide a legal citation or describe how every relevant national law or contract would apply, so the scope of that claim cannot be established from the announcement alone.
Meyer also says model-provider contracts quoted in the book can cap liability at roughly a year of fees. The specific providers, contract terms and circumstances are not given in the supplied material. It remains unclear how courts would allocate responsibility in a dispute involving an autonomous agent, its user, the deploying organization and its vendor. The steel purchase is a composite example, not evidence of an actual loss, and the source does not report a real-world case involving that scenario.
How Organizations Put Controls in Place
The ebook is available in its first edition as of October 2026, with sales links listed for Apple Books, Barnes & Noble, Kobo, Thalia, Bookshop.org, Smashwords, Fable and Vivlio. Meyer also describes a five-minute video about the book, narrated using an AI voice clone of the author.
The next test for the issues raised is how organizations and lawmakers handle actual agent-driven disputes: who had authority, what information the agent received, whether a person could realistically review or reject the action, and what remedies apply. The source does not announce a follow-up legal decision or policy milestone. For now, Meyer’s proposed controls are recommendations for organizations to consider, not a reported change in law.
Key Questions
What is Many Hands about?
Thorsten Meyer’s ebook examines who pays, who must explain and fix a failure, and who bears moral responsibility when an AI agent acts for a person or organization.
Is the steel-purchase story a real incident?
No. Meyer describes Theresa and the purchasing decision as a composite scenario built from common situations. The source does not identify it as an actual company’s loss.
What safeguards does Meyer recommend?
He recommends naming a responsible person for each agent, limiting permissions to the task, keeping decision records, providing a route for affected people to raise concerns, and using records to improve systems.
Does the book establish who is legally liable for an agent’s actions?
No. It examines the question but does not report a court ruling that settles liability for AI agents generally. The outcome of a dispute would depend on the applicable law, contracts and facts.
When was the ebook published?
The source dates the ebook’s publication and first edition to October 2026, with the announcement published October 27.
Source: ThorstenMeyerAI.com
