iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
There is no single person automatically responsible when AI causes harm. The answer depends on the jurisdiction, the kind of harm, the applicable legal claim, and what the people and organisations involved did. A developer or provider, the organisation that deployed the system, a professional or individual who used its output, or another party may be relevant. The AI system’s output alone does not settle who is legally responsible.
What does “responsible” mean?
People often use “responsible” to mean either “required to follow the rules” or “liable to compensate someone for harm.” Those are related, but separate, questions. A regulator may investigate whether an organisation complied with AI rules; a civil claim asks whether a party is liable under the law that governs the particular loss. A regulatory breach may be relevant to a civil case, but it does not automatically determine who must pay damages.
Nor does describing an AI system as having made a decision make the system itself a legal person. The practical inquiry is what role each human or organisation played, what conduct or product condition is alleged to have caused the harm, and what law applies.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Which people or organisations might be involved?
Responsibility is not assigned by a universal ranking of “the developer” versus “the user.” These roles are useful starting points for identifying whose conduct or product may need to be examined; they do not, by themselves, establish liability.
#1 Best Overall
| Role | What to examine |
|---|---|
| Developer, provider, or product manufacturer | How the system or software was designed and supplied, whether the claim concerns a defective product, and whether that alleged defect caused the harm. |
| Deploying organisation or employer | Why it selected the system, how it configured and monitored it, and whether required human oversight or other applicable duties were met. |
| Professional or other user | How the person used the system, whether they checked or acted on its output appropriately in context, and whether their conduct contributed to the harm. |
| Person who relied on the output | What the system represented, how the person relied on it, what loss followed, and whether a legal route applies to that relationship and harm. |
| Other parties | Whether another organisation or person supplied, integrated, controlled, or acted on the system in a way that matters under the governing law. |
How the EU rules illustrate the distinction
The EU provides a current example of why regulatory duties and compensation claims should not be conflated. The AI Act assigns compliance duties to regulated parties, including providers and deployers, and gives the AI Office and national market surveillance authorities supervisory and enforcement roles. For high-risk systems within the Act’s scope, deployers have duties that include assigning competent human oversight and monitoring operation. The regulation states in Article 14(4): “Deployers shall assign human oversight to natural persons who have the necessary competence, training and authority, as well as the necessary support.” That is a compliance requirement for the covered systems; it is not, by itself, a rule making a deployer liable for every harmful output.
Product liability for software and AI systems
Directive (EU) 2024/2853 expressly includes software, including AI systems, in the EU product-liability framework and treats a developer or producer of software, including an AI-system provider, as a manufacturer. This route concerns damage caused by a defective product; it is not a universal compensation rule for every harmful answer, service, or use of AI. The Directive applies from 9 December 2026, subject to its terms, temporal scope, and national implementation. Because that date is still in the future as of 9 October 2026, do not assume the revised rules govern an event that occurred before they apply.
Other civil claims and the withdrawn proposal
The revised Product Liability Directive does not erase other possible routes. Depending on the facts and local law, a person may also have a claim under contract law or non-contractual national rules outside a manufacturer’s product-liability claim.
The European Commission’s proposed AI Liability Directive was intended to improve proof in certain non-contractual civil claims involving AI. It was a proposal, not an enacted directive, and EUR-Lex records its withdrawal on 6 October 2025. Its proposed procedures should not be presented as a remedy currently in force.
Rank #3
What facts can determine who is liable?
There is no jurisdiction-independent checklist that decides every case. The following questions help identify the issues that may need to be resolved under the applicable law:
- What happened? Identify the system’s role in the decision or event and the specific harm alleged.
- Which parties controlled what? Establish who developed or supplied the system, who selected and configured it, who monitored it, and who acted on its output.
- What is the alleged legal wrong? It might concern a defective product, a failure to meet a regulatory duty, a contractual obligation, negligent human conduct, or another legally relevant act or omission. These are possible lines of inquiry, not findings that a particular party breached a duty.
- Did the alleged defect or conduct cause the harm? A claim generally needs to connect the alleged cause to the loss under the rules governing that claim; the mere presence of AI in a process does not establish that connection.
- What evidence exists? Relevant material may include system and configuration records, instructions, monitoring and review records, communications, contracts, and evidence of the harm. Which evidence matters and how it can be obtained depend on applicable law and procedure.
- Where and when did it happen? The governing jurisdiction, the event date, local implementation, and any relevant contract can change which rules apply.
Opacity and difficulty tracing an AI-related decision can make it harder to identify a potentially liable party and prove a claim. That difficulty does not itself establish liability or create a special procedure for obtaining compensation.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What should someone do after AI-related harm?
If a real incident has caused loss, focus first on preserving the facts rather than deciding who is to blame. Record the date, the system and product involved, what the system produced or did, who reviewed or acted on it, and what harm followed. Keep relevant communications, contracts, notices, and other records where you can lawfully do so. Avoid altering or deleting material that may help explain what happened.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Repair Windows errors before they cause bigger problems3Scan for outdated or missing drivers - takes under a minuteThen identify the location and date of the event, the parties’ roles, the alleged defect or conduct, and the loss. A qualified lawyer in the relevant jurisdiction can assess possible claims and deadlines against current local law. This article provides a general framework, not a determination of liability or legal advice for a particular incident.
Can I sue a company if its AI harms me?
Possibly, but harm involving AI does not by itself establish a valid claim against a company. Whether a claim is available depends on the jurisdiction, the facts, the company’s role, the applicable legal route, causation, and evidence. In the EU, product liability may be relevant when the claim concerns damage caused by a defective product within the applicable rules; contract and other national civil-law routes may also matter.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
