¾«¶«´«Ã½

Are Ai Instant Answers Accurate?

Since the release of publicly available artificial intelligence (Ai) tools such as ChaGpt by OpenAi in 2022 and Claude by Anthropic in 2023, people increasingly ask whether they need a lawyer at all. Having access to an instant, low-cost tool instead of taking an appointment and paying to get legal advice seems both interesting and cheap. However, Ai can ‘hallucinate’, i.e., generate false and misleading information, while presenting it confidently as a real fact.  The Judiciary have also reported that AI can often mislead clients about their prospects of success in litigation or about the amount that their claim is worth. This can often mean that client’s expectations can be inflated meaning that mediation and settlement can be difficult.

According to Dominick Romano, founder of the Ai management platform drainpipe.io, several categories of hallucinations can be listed:

• Factual inaccuracies (incorrect facts or dates);
• Fabricated content (fictional entities and sources);
• Nonsensical outputs (absurd or irrelevant responses); and
• Over-generalizations (overly broad or simplistic answers).

Asking Ai About the Law.

Ai has access only to the documents provided by the user, and it can make mistakes, even when searching the Internet. It also needs to be understood that those types of language models don’t know what is ‘true’ and lack common sense. Instead, they function on a probabilistic system, which will predict the most probable next word. The model could try and simplify legal processes but could misinterpret an important step or suggest a step which is not relevant for your case. Ai may be able to tell you what the law is, but it may not be able to tell you what you should do about it. Poor and uninformed use of Ai can lead to poor results and consequential problems, especially in complicated and dynamic or nuanced situations.

Dealing with a case at court requires much more than knowing the applicable law: being strategic, subtle, and adaptable are also a lawyer’s core competencies. Since an Ai has been trained on huge quantities of text, its perspective and contextualization aren’t drawn from professional experiences. It never faces a case that goes wrong, and it doesn’t have to answer to a superior, competitors on the work market, or a reputation to lose.

Therefore, you should:

• Avoid providing personal information in Ai prompts;
• Check any laws or court decisions with a qualified solicitor;
• Treat any compensation amount as provisional and have it validated by a solicitor or a legal advice clinic.

UK Lacks Safeguards Against Ai Downsides.

Since UK legislation and litigation cases are very specific (e.g., procedural expectations, evidential thresholds), it seems very risky to entrust a litigation case to a digital bot trained on international data. Even qualified solicitors have encountered problems when using it. For example, in R (Ayinde) v London Borough of Haringey; Al-Haroun v Qatar National Bank Qpsc [2025] Ewhc 1383 the witness statement of the claimant proposed 18 court rulings that didn’t exist. Solicitors have been reprimanded (and even struck off) for failing to verify Ai-generated research. This could constitute negligence.

How ¾«¶«´«Ã½ Legal Advice Clinic (BULAC) Can Help.

Legal disputes are complex, emotionally draining, and costly. This is why we can help you in civil law matters by offering initial guidance to help you understand your options. Indeed, BULAC can:

• Advise you on the accuracy and relevance of the answers provided by Ai;
• Provide you with a genuine, tailored analysis of your situation; and
• Offer you an experienced, human perspective.

At ¾«¶«´«Ã½ Legal Advice Clinic  students provide free, confidential advice to the local community under the supervision of a qualified solicitor. You can make an appointment by contacting 01248-388-411 or email bulac@bangor.ac.uk.