Artificial Intelligence (AI) has moved from science fiction into everyday life remarkably quickly. Businesses use it to draft emails, write reports, analyse data and deal with customers. Individuals are using it to prepare letters, contracts and even court documents.
Used properly, AI can be an extremely useful tool. Used carelessly, however, it can create significant legal and financial problems.
In this article, Sophie Shilling Solicitor of Kingsfords Solicitors Litigation and Disputes Team with local offices in Ashford, Cranbrook and Hythe provides a valuable insight to this rapidly evolving area of law. It covers what’s happening, why it matters, and the strategic & practical implications for clients and advisers alike.
We look at two questions we are increasingly asked by clients:
- Who is liable if an AI system causes commercial loss or damage?
- Can I use AI to draft my own court documents?
Who is Liable if an AI System Causes Commercial Loss or Damage?
One of the most common misconceptions about AI is that it somehow takes responsibility for its own mistakes.
Unfortunately, the law does not work that way.
If a member of staff makes an error, the employer may be liable. If a computer system makes an error, the business operating that system may still be liable. Courts are generally interested in who deployed the technology and who benefitted from its use, rather than whether the mistake was made by a human or a machine.
In the UK there is currently no separate legal personality for AI. An AI system cannot be sued in its own right, in the same way that a company or an individual can. Responsibility will usually fall on the business, organisation or individual using the technology.
Real-World Examples
The risks are far from theoretical.
One well-publicised case involved Air Canada. Its customer service chatbot provided incorrect information regarding bereavement fares to a customer. The customer relied upon that information when booking flights. When the airline later refused to honour what the chatbot had said, legal proceedings followed. The British Columbia Civil Resolution Tribunal in Canada found in favour of the customer and rejected arguments that the chatbot was somehow responsible for its own statements. The information provided by the chatbot was treated as information provided by the airline itself.
Businesses have also reported losses caused by AI systems generating inaccurate information, making incorrect recommendations or producing misleading content. While AI can process huge amounts of information quickly, it can also be confidently wrong! Anyone who has used AI for more than a few minutes will probably have discovered that it occasionally invents facts with impressive confidence.
This creates obvious risks where AI is being used in customer service, compliance, financial decision-making or contract management.
What Does the Law Expect?
The legal position is still developing, but certain principles are already clear. Businesses are expected to act reasonably.
That means implementing proper safeguards and supervision. AI should generally be treated like a very clever junior employee. It may be capable, fast and enthusiastic, but it still needs supervision, and somebody should check its work before important decisions are made.
Appropriate safeguards may include:
- Human review of important outputs.
- Clear internal policies governing AI use.
- Staff training.
- Testing and monitoring AI systems.
- Maintaining records of decisions and outputs.
- Ensuring compliance with data protection obligations.
A business that simply switches on an AI system and assumes everything will be fine may find itself facing difficult questions if something goes wrong.
The key point is simple: using AI does not remove legal responsibility. In many cases, it simply changes the way that responsibility arises.
Can I Use AI to Draft My Own Court Documents?
The short answer is “yes”.
The better question is whether you should.
AI can certainly help individuals understand legal issues, organise information and produce first drafts of documents. For many people, it can be a useful starting point.
However, there is a considerable difference between producing a document and producing a document that is legally accurate, procedurally compliant and persuasive.
Court proceedings are governed by rules, deadlines and legal principles that are often far more complicated than they first appear. A document that looks impressive may still contain serious legal defects.
The Problem of AI Hallucinations
One of the most significant risks is the phenomenon known as "hallucination".
This occurs when an AI system generates information that sounds entirely convincing, but actually is completely false.
In legal work, this can be particularly dangerous.
There have now been several reported cases in which lawyers themselves have relied upon fictitious legal authorities generated by AI. In the widely reported case of Mata v Avianca, in the US, lawyers filed court documents containing non-existent cases generated by ChatGPT. The court imposed, sanctions after discovering that the authorities cited simply did not exist.
The issue has not disappeared or is confined to the US. UK Courts and regulators continue to report instances of fabricated authorities and inaccurate legal references appearing in legal submissions prepared with AI assistance. Regulators have repeatedly emphasised that lawyers remain responsible for checking the accuracy of any AI-generated material before it is filed with the court. A few reported cases to-date include Ayinde v London Borough of Haringey, that of Al-Haroun v Qatar National Bank, and also in 2025 HMRC V Gunnarsson.
Academic research has also demonstrated that even specialist legal AI tools can generate inaccurate legal information and fictitious authorities.
What Could Go Wrong?
If court documents contain inaccurate legal arguments, procedural errors or fictitious authorities, the consequences can be serious.
Depending on the circumstances, a court may:
- Reject the document.
- Strike out part of a claim or defence.
- Order amendments.
- Award costs against a party.
- In extreme cases, impose sanctions.
Perhaps more importantly, a weak or inaccurate document can damage a party's credibility before the court.
Judges generally expect parties to ensure that documents filed with the court are accurate, whether they were drafted by a solicitor, a barrister, a litigant in person or a computer.
So Should I Use AI?
AI can be a useful assistant.
It can help identify issues, suggest wording and save time.
What it should not be treated as is a substitute for legal advice.
To put it another way, using AI to conduct litigation can sometimes resemble performing your own dental treatment after watching a few online videos. You may save money initially, but there is a risk that somebody with professional qualifications will become involved eventually.
As solicitors, we must confess that we may have a slight conflict of interest when making that observation. We are, after all, the people whose services you would be engaging.
However, experience suggests that many clients come to us not because they lacked confidence, but because they discovered that litigation is rather more complicated than it first appeared.
AI is undoubtedly changing the legal landscape, and it will continue to do so. It offers enormous opportunities for businesses and individuals alike. Nevertheless, it remains a tool rather than a replacement for human judgment.
Whether you are deploying AI within your business or considering using it to prepare court documents, the safest approach remains the same: trust it enough to help, but not enough to stop checking its work.
Contact Us
If you are an entity or party considering using Artificial Intelligence in your client facing dealings, or have a potential or existing litigation dispute and wish to discuss any of Artificial Intelligence issues raised, Sophie Shilling and other Solicitors of Kingsfords Litigation and Disputes Team are on hand to support you. Please contact your local Kingsfords office in Ashford, Cranbrook or Hythe. Alternatively, you can click on the “Get In Touch” icon on the right hand side of our Website to send an enquiry and a member of our team will get back to you promptly. You can also call us directly on 01233 665544, where we will be happy to discuss your concerns without obligation.