Skip to content 
AI is no longer a distant concept, and is quickly becoming a key figure in the legal world. It has been integrated into law firms and courtrooms across the globe. Whether it is used to: sift through large volumes of legal documents, draft witness statements, or assist with enhanced legal research, it has had a significant impact on the legal field. Technologies like this offer incredible advantages, which enhance the provision of legal services and help firms gain a competitive edge.
The increase in the use of AI is evident within the legal field, helping speed and capacity, saving costs and structuring their business in new ways. The Solicitors Regulation Authority’s outlook report from 2023 explores how the use of AI is rising rapidly. At the end of last year, the number of the largest solicitors’ firms who were using AI doubled from just three years ago. It was also found that 72% of financial service firms used AI. As of 2025, statistics from a Thomson Reuters survey also show that 77% of law firm professionals are using Gen AI for reviewing documents, with 77% of all legal professionals using it for summarising documents.
However, alongside this wave of innovation, there comes complex legal and ethical concerns.
The SRA’s Chief executive, Paul Philips, stated: “Firms need to make sure they understand and mitigate against them – just as a solicitor should always appropriately supervise a more junior employee, they should be overseeing the use of AI. They must make sure AI is helping them deliver legal services to the high standards their clients expect.”
Some of the most apparent risks which are inherent with the use of AI are:
Accountability
- Law firms are to be held accountable to clients for any legal services provided, regardless of whether AI systems have been used or not.
- The issue of ‘duty of care’ and how a breach will be assessed is becoming an increasing concern when AI is of use.
- Causation may also be difficult if it is hard to identify how a failure occurred.
- The unpredictability of AI may also contribute to the difficulty of claimants to prove if a loss was foreseeable.
Accuracy
- It is within the nature of humans to make mistakes, especially those who are new to the legal field and need supervision.
- Just like humans, AI needs this supervision as it may not always be accurate.
- However, accuracy cannot always be measured on a linear scale, and without more research into the accuracy of AI, we cannot guarantee its reliability. There is increased risk when firms put their trust in AI, which can be easily overlooked.
Bias
- The ‘luddites’ argue that AI creates answers which are prone to bias, following incorrect patterns from the data.
- These biases can be subtle and can result in outcomes which are not entirely fair or correct, leading to miscarriages of justice or discrimination which risks undermining the rule of law.
Confidentiality
- Generally, law firms tend to deal with highly sensitive and confidential information which should be protected.
- Questions need to be asked before entering any information like: where is this information going to be stored; what will happen to this information; what is the agreement with the client regarding the confidentiality of their information?
Real world examples
These risks can be shown through the recent case against Qatar National Bank where the claimants made 45 case-law citations, 18 of which turned out to be fictitious. The claimant admitted that they may have used publicly available AI tools on Google or Safari. This resulted in legal action for wasted legal costs and ultimately the pupil barrister and the law centre were found negligent. Dame Victoria Sharp, president of the King’s bench division said there were “serious implications for the administration of justice and public confidence in the justice system if artificial intelligence is misused”. Misuse or being overly reliant on AI could mean that solicitors and barristers are at increased risk of being negligent.
Similarly, the US firm, despite its praised experience in defending prison cases, Butler Snow is facing sanctions by the federal judge after an attorney at the firm cited cases generated by artificial intelligence – which turned out not to exist. A few of the wide range of sanctions include fines, mandated continuing legal education and temporary suspension. The attorney stated that he “failed to verify the case citations returned by ChatGPT through an independent review in Westlaw before including them”. Therefore, over-reliance on inaccurate AI tools have meant harsh sanctions and increased findings of negligence for professional and reputable law firms.
Despite these setbacks, the Solicitors Regulation Authority announced it had authorised Garfield Law ltd to provide regulated legal services in England and Wales. This marks an inflection point for the evolution of regulated legal services in the UK. Considering the rapid progress of AI and its increased use, it would be reasonable to not completely ban or ignore it, and instead, regulate it. The service Garfield offers, though highly automated, requires client approval and is supervised by a qualified litigator. Common AI errors like the ‘hallucinated’ case law, as we have seen in the case against Qatar National Bank, will then be easily avoided. This is a remarkable step towards integrating reliable and more accurate use of AI to overall increase efficiency, and affordability, for law firms. Implementing more services like this will allow more firms to engage with the innovative technology and assistance AI has to offer.
To Conclude
Overall, AI can be accurate, but it can never be perfect. Ultimately in most areas of law, one small mistake can lead to detrimental consequences. Human oversight is therefore essential when regulating AI because replacing human judgement in high stake cases can open doors for an array of issues within our justice system. Therefore, as much as AI can be used as an innovative tool to assist individuals with basic tasks, it should never be used as a complete replacement for human judgement. Regulation of AI is in progress, but there is still emphasis on accountability. Solicitors remain reliable for outputs and consumer protections are still preserved.
Hannah Jamil on work experience at Lawdit Solicitors
