Artificial intelligence is becoming part of everyday legal practice, assisting with research, document review, drafting and analysis. But what happens when the technology does more than make an obvious mistake?
An invented court judgment can eventually be exposed. A more difficult problem arises where AI identifies genuine law but recommends the wrong argument, overlooks the stronger claim or produces convincing analysis that sends a case down the wrong path.
The Solicitors Regulation Authority (SRA) received 42 reports relating to potential AI misuse between July 2025 and July 2026, with ongoing investigations involving inaccurate legal citations, supervision and confidentiality. The regulator is clear that using AI does not reduce a solicitor's responsibility for the work and advice provided.
This raises an increasingly important question for solicitor negligence claims: when does using AI stop being a useful tool and begin to undermine the professional judgement a client was paying for?
Key Takeaways
- Using AI does not itself amount to solicitor negligence.
- Solicitors remain responsible for checking AI-assisted research, analysis and advice.
- AI can cause problems without hallucinating, including selecting a weak legal argument or overlooking a stronger claim.
- A claimant must establish that the solicitor's failure caused a financial loss, which in lost litigation cases may be the loss of a real and substantial chance of a better outcome.
- Keep legal advice, correspondence and the solicitor's file, and investigate potential claims promptly because limitation periods apply.
“The very rapid development of AI is becoming a significant factor in solicitor negligence and that will continue to be the case. The legal analysis is still developing through the courts but it is very likely that this will be an area where claimants will have legitimate claims arising. This may prove to be the case in other areas of professions involving accountants, engineers, architects and so on where AI is also becoming increasingly important.” Tony Hill - Head of Professional Negligence & Commercial Litigation
What Counts as Negligent Use of AI by a Solicitor?
A solicitor using ChatGPT or another AI system has not automatically acted negligently. AI can legitimately assist lawyers with research, drafting, document analysis and administrative work.
The question is whether the solicitor exercised the reasonable skill and care expected of a competent practitioner when using it.
The SRA's August 2026 warning notice states that existing professional obligations continue regardless of the technology used. Solicitors remain responsible for providing competent services, supervising work and ensuring information given to clients, courts and third parties is accurate.
The Law Society takes a similar approach. Its October 2026 AI practice note states that generative AI cannot reliably check its own accuracy and that outputs require appropriate human verification, oversight and professional judgement. It also identifies potential professional liability where AI is used inappropriately, an unsuitable system is selected or its output is insufficiently tested or validated.
A poor result alone is therefore insufficient. The issue is whether the solicitor relied on AI in circumstances where competent independent analysis should have identified or prevented the error.
Concerned AI or poor legal advice may have affected your case?
Our professional negligence team can review your solicitor's advice, litigation strategy and case file to assess whether the service you received may support a claim.
Common Examples of AI-Related Solicitor Negligence
The best-known AI failure is the hallucination: information that appears convincing but is false. Legal negligence risks, however, extend considerably further.
Hallucinated Cases and Authorities
AI can generate fictitious judgments, quotations or legislation with realistic-looking citations.
In Ayinde v London Borough of Haringey and Al-Haroun v Qatar National Bank [2025] EWHC 1383 (Admin), the High Court addressed actual or suspected use of generative AI involving false legal material. The court emphasised that lawyers using AI-assisted research have a professional duty to check it against authoritative sources before relying on it.
Submitting an invented case is an obvious failure. The more difficult problems arise when everything appears genuine.
Choosing the Wrong Legal Argument
AI may identify real authorities and accurately describe parts of the law, yet still recommend the wrong route.
For example, a commercial dispute might support several possible claims. AI could focus the solicitor on one contractual argument while overlooking a materially stronger case based on misrepresentation, professional negligence or another cause of action.
Nothing has been invented. The problem is the legal strategy.
Solicitors are instructed to apply judgement to the client's particular facts, evidence and objectives. If an AI-generated analysis becomes the foundation of a case without sufficient independent consideration, a plausible answer can potentially be more damaging than an obvious hallucination.
The Law Society specifically advises solicitors to identify omissions, unsupported propositions, bias and over-confident conclusions, rather than accepting AI outputs as definitive.
Missing the Stronger Claim or Important Evidence
AI can also produce a detailed analysis while omitting the point that matters most.
That could mean overlooking:
- another potential defendant;
- an important defence or remedy;
- a contractual clause that changes the legal position;
- evidence contradicting a key witness;
- the need for specialist or expert evidence; or
- a limitation or procedural issue.
This risk becomes particularly significant when AI is used to review large quantities of documents. Summarising information is not the same as recognising its legal significance.
A solicitor who receives an apparently comprehensive summary must still determine whether the evidence has been properly understood and whether anything material is missing.
Correct Law, Wrong Advice
An AI system does not need to get the law wrong to produce poor legal advice.
It might correctly explain the test established by a judgment but incorrectly conclude that the client's evidence satisfies it. Alternatively, it could overstate the prospects of success, leading a client into expensive litigation, or undervalue a viable claim and encourage an unnecessarily low settlement.
The danger is that an authoritative tone can make uncertain conclusions appear settled.
This is closely connected with automation bias: the tendency to defer excessively to automated output. The Law Society identifies this as a significant risk in legal practice because professional responsibility remains with the lawyer even where AI contributed to the conclusion.
Who Is Responsible When a Solicitor Uses AI?
The solicitor cannot simply transfer responsibility to the software.
The SRA states that AI has no separate legal personality and that regulated lawyers remain accountable for their work regardless of how it was prepared. This applies to work produced personally and work carried out by people they supervise.
Supervision is particularly important.
A trainee or paralegal can now generate substantial research or drafting in minutes. A supervising solicitor who assumes that work has been properly researched, without understanding how AI was used or checking important conclusions, could allow errors to reach the client or court.
The Upper Tribunal considered similar concerns in UK and R (Munir) v Secretary of State for the Home Department [2026] UKUT 81 (IAC). It stressed that legal professionals must ensure arguments presented to tribunals are factually and legally accurate.
AI also creates a separate confidentiality risk. The SRA warns that entering client-sensitive material into AI systems without appropriate safeguards may compromise confidentiality, data protection and potentially legal professional privilege.
Proving an AI-Related Solicitor Negligence Claim
Evidence of AI use is not enough to establish a legal professional negligence claim.
The usual principles still apply. A claimant will generally need to establish:
- Duty of care: the responsibilities the solicitor undertook.
- Breach: work that fell below the standard expected of a reasonably competent solicitor.
- Causation: what would have happened if competent advice had been given.
- Loss: the financial difference caused by that failure.
The causation question can be particularly important where the complaint concerns legal strategy.
Suppose AI led a solicitor towards a weak argument while a stronger claim was never pursued. It is not enough simply to identify the alternative argument afterwards. The claimant will usually need to establish that competent legal advice would have identified it, and that they lost a real and substantial, rather than merely speculative, chance of achieving a better outcome.
They do not need to prove that the alternative claim would definitely or probably have succeeded. Instead, compensation is usually assessed by reference to the value of the chance that was lost.
Relevant evidence may include the engagement letter, legal advice, pleadings, emails, attendance notes, settlement recommendations and the solicitor's complete file. Where available and relevant, records showing how AI-assisted analysis contributed to a decision may also become important.
Independent expert evidence can sometimes be required to assess what a reasonably competent practitioner should have done.
What You Can Recover
Compensation in solicitor negligence claims generally aims to put the claimant in the financial position they would have occupied had competent legal services been provided.
Potential losses may include:
| Type of loss | What needs to be established |
|---|---|
| Lost underlying claim | Competent handling would have given a real and substantial prospect of successfully pursuing the claim, with compensation valued according to that chance |
| Reduced settlement | Negligent advice caused the client to accept less than the claim was reasonably worth |
| Unnecessary litigation costs | Costs were incurred because an inappropriate strategy was pursued |
| Remedial professional fees | Further reasonable expenditure was required to correct the solicitor's mistake |
For example, AI-assisted analysis might incorrectly assess a claim as weak and contribute to a settlement substantially below its proper value. The claimant would still need evidence establishing what competent advice should have been and that it would have given them a real and substantial chance of a better result.
Likewise, unsuccessful litigation is not automatically evidence of negligence. Courts regularly have to decide genuinely uncertain legal questions. The issue is whether the solicitor's conduct fell below the required professional standard, rather than simply whether the case was lost.
The Time Limits That Apply
The use of AI does not create a separate limitation period.
In England and Wales, claims founded on simple contract generally must be brought within six years of the breach of contract. Claims in tort are also generally subject to a six-year limitation period, which runs from when damage is first suffered. In claims against solicitors, damage is often suffered at the time the negligent advice is acted upon, which may be long before the client realises anything has gone wrong.
For negligence claims brought in tort, other than personal injury claims, involving latent damage, section 14A of the Limitation Act 1980 can provide a three-year period from the required date of knowledge where that expires later. Section 14A does not apply to claims brought in contract. A fifteen-year longstop generally applies under section 14B, running from the date of the negligent act or omission, subject to applicable exceptions such as deliberate concealment.
AI-related errors could be difficult to discover immediately. Advice may appear convincing until another solicitor reviews the case and identifies an argument or piece of evidence that was overlooked.
Discovering that AI was used does not itself determine when limitation starts. Obtain advice on the particular circumstances rather than waiting for the underlying dispute to conclude.
How an AI-Related Solicitor Negligence Claim Proceeds
The first stage is normally to obtain the solicitor's file and independently review what happened.
The investigation should identify the actual professional failure rather than treating AI use itself as the complaint. That may involve comparing the strategy pursued with the legal arguments, evidence and options that were reasonably available at the time.
Where grounds for a claim exist, a detailed Letter of Claim can be sent under the Pre-Action Protocol for Professional Negligence, explaining what the solicitor allegedly did wrong, what should have been done and how the failure caused the claimed loss.
The parties can then investigate, exchange relevant information and explore settlement or mediation. Court proceedings may become necessary if liability, causation or the value of the loss remains disputed.
FAQs
Can I sue my solicitor for using AI?
Using AI does not itself create a claim. You may have grounds for a professional negligence claim where your solicitor fell below the standard expected of a reasonably competent practitioner and that failure caused financial loss.
Is my solicitor responsible if AI gives them incorrect information?
Solicitors remain responsible for the legal work and advice they provide. The SRA states that using AI does not reduce, remove or transfer their professional responsibilities.
Can choosing the wrong legal argument amount to solicitor negligence?
Potentially. Choosing an unsuccessful argument does not automatically establish negligence because litigation involves professional judgement. A claim may arise where the strategy fell below the standard expected of a competent practitioner and pursuing it caused financial loss.
What if the AI used genuine cases but reached the wrong conclusion?
A genuine citation does not make the overall advice correct. Solicitors must consider whether the authorities support the argument, whether important law or evidence has been omitted and whether the conclusion properly applies to the client's circumstances.
How long do I have to bring an AI-related solicitor negligence claim?
Many professional negligence claims involve a six-year limitation period, but the starting point can differ between contract and tort claims. For claims in negligence (tort), latent damage provisions may extend the period in some circumstances, subject to a 15-year longstop. Obtain advice promptly because limitation is highly fact-specific.