Skip to content
Worthclock

Blog / 8 min read

Who Is Liable When AI Gets a Legal Answer Wrong: the Lawyer, the Firm or the Vendor?

By Shang Ahmed · 15 November 2026

When AI gives a wrong legal answer, who is responsible? Why liability usually stays with the lawyer and firm, and how to protect yourself.

Quick Answer

In almost every case, responsibility stays with the lawyer and the firm, not the AI vendor. Regulators and courts treat AI as a tool, and the professional who relies on it remains accountable for the advice and the work. The SRA has said individuals remain responsible for the work they produce whether or not AI is used, and courts have sanctioned lawyers for filing AI generated fake citations. Vendors usually limit their liability in their contracts. The practical protection is a review process, clear supervision and insurance that covers AI assisted work.

What Does "AI Liability" Mean for a Law Firm?

For a law firm, AI liability means who has to answer for the consequences when AI assisted work turns out to be wrong. That can include regulatory action, court sanctions, a negligence claim from a client or a dispute with the AI provider. Each has a different answer, but they mostly lead back to the same place.

Is the Lawyer Responsible?

Yes. The lawyer who relies on AI output remains professionally responsible for it, exactly as if they had produced it themselves. Using AI does not lower the standard of care or move the duty onto the tool.

The SRA put this plainly when it published its warning notice on the misuse of AI in August 2026: individuals remain responsible for the work they produce and the advice they provide, regardless of whether AI is used, and the use of AI does not change the professional standards expected of solicitors and law firms.

Courts take the same view. In Mata v. Avianca (S.D.N.Y. 2023), a New York federal court fined two lawyers and their firm $5,000 after they filed a brief citing cases invented by ChatGPT. In Ayinde v London Borough of Haringey [2025] EWHC 1383 (Admin), the High Court in London warned that lawyers who put false citations before a court may face referral to their regulator, contempt proceedings and, in the most serious cases, criminal investigation. The same pattern is covered in AI hallucinations in legal filings.

The consequences have grown since. In 2026, two attorneys in Oregon were assessed $110,000 in fines and fees over hallucinated citations, and the Nebraska Supreme Court indefinitely suspended a lawyer after 57 of the 63 citations in his brief turned out to be faulty or nonexistent.

Is the Firm Responsible?

Yes. The firm is responsible for supervision, systems and controls, and it will usually be the defendant in a client's claim. A firm cannot avoid responsibility by pointing to an individual's choice of tool.

The SRA's warning notice reminds firms that supervisors remain accountable for work carried out through others, and that firms need effective governance structures, systems and controls to manage AI risks. The court in Ayinde also stressed that people with leadership responsibilities must take practical steps to make sure everyone in their organisation understands and follows their professional obligations when using AI.

In the US, ABA Formal Opinion 512 makes a similar point: partners and managers should set clear policies on permitted AI use, and supervisors must make sure lawyers and staff are trained to follow them.

In practice, a client who receives wrong advice will usually bring a claim against the firm, and the firm's professional indemnity insurance is where that claim ends up.

Is the AI Vendor Responsible?

Rarely, at least in a way that helps the firm. Most AI providers limit their liability in their contracts, disclaim the accuracy of outputs and make the customer responsible for reviewing results.

Firms sometimes assume new EU rules will change this. They are unlikely to help in the typical case:

  • The AI Liability Directive will not happen for now. The European Commission announced in February 2025 that it was withdrawing the proposal, saying no foreseeable agreement had been reached.
  • The new Product Liability Directive is aimed at defective products. Directive (EU) 2024/2853 brings software, including AI, within strict product liability for products placed on the market from 9 December 2026. But it focuses on harm such as personal injury, damage to property and loss of data not used for professional purposes. It is not designed to compensate a firm for a wrong legal answer that a lawyer relied on.

That leaves the contract. Before approving a tool, firms should read what the provider actually promises about accuracy, security and liability, and should assume the answer on accuracy will be "not much".

Does a Client's Consent Shift Responsibility?

No. A client agreeing that the firm may use AI does not move responsibility for the quality of the work. Consent addresses confidentiality and transparency. It is not a waiver of the duty to give competent advice.

How Can a Firm Protect Itself?

A firm protects itself with a review process, clear supervision, careful tool selection, honest records and the right insurance. None of these are expensive, and together they make AI errors much less likely to reach a client or a court.

  1. Require human review of all AI assisted work before it leaves the firm.
  2. Check every citation and quotation against the original source, without exception.
  3. Approve tools deliberately, reading the provider's terms on accuracy, data and liability.
  4. Supervise how work is produced, not just the final result, especially for trainees and junior staff.
  5. Train everyone on where AI tools make mistakes and how to spot them.
  6. Keep records of the firm's AI policy, approved tools and training, so it can show reasonable steps were taken.
  7. Talk to your insurer about how your professional indemnity cover treats AI assisted work, and tell them about your controls. See also whether underwriters will ask about your firm's AI use.

FAQ

Can a lawyer blame the AI tool for a mistake?

No. Regulators and courts treat AI as a tool. The lawyer remains responsible for checking the output and for the advice given.

Will professional indemnity insurance cover AI errors?

Often, because the claim is about the quality of the legal work. But policies differ, so firms should confirm with their insurer how AI assisted work is treated.

Can we sue the AI vendor if its tool gives a wrong answer?

In most cases the contract will limit or exclude that. The new EU Product Liability Directive is aimed at harm such as personal injury and property damage, not wrong professional advice.

Are trainees personally liable if they use AI badly?

Trainees have their own professional obligations, but supervisors and the firm remain accountable for work carried out through them.

Related reading

The Bottom Line

When AI gets a legal answer wrong, the responsibility almost always stays with the lawyer and the firm. Vendors protect themselves in their contracts, and new EU liability rules do not change that for everyday legal work. The firms that are protected are the ones that review every output, supervise properly and can show the steps they took.

Want to know where AI errors could slip through in your firm's workflows? Our AI Tools Assessment maps how AI is used and where review steps are missing.

Book Your Assessment

Sources