People

Careers

Corporate

Commercial

Employment

News & Insights

Blog

News

Videos

Who we are

Environment & Sustainability

Diversity, Equity & Inclusion

WONDER Foundation

Get legal updates and practical guidance straight to your inbox.

You’ve successfully signed up to the Linnear Legal mailing list

Home

Resources

<p>Registered in England &amp; Wales at 73 Shelton Street, Covent Garden, London, WC2H 9JQ</p>
<p><a href="tel:+442081596000">Tel: +44 (0)20 8159 6000</a></p>
<p>Company Nr: 15894546</p>
<p>ICO Registration Nr:  ZB842887</p>
<p style="text-align: left;">VAT Registration Nr: 484 5226 73</p>
<p>Linnear Legal is an unregulated legal services provider, meaning the firm itself is not regulated by the Solicitors Regulation Authority (SRA). However, the solicitors employed by the firm are individually regulated by the SRA. Please see our Terms and Conditions for further information on our regulatory status, complaints procedure and insurance arrangements.</p>


Terms & Conditions

Cookie Policy

Privacy Policy

Services

Insights

About

We are driven by doing things differently, from every angle. 

Taking our Carbon Neutral Certified  status seriously.  

Putting equity and inclusion at the heart of everything we do.

Supporting sustainable employment and gender equality with WONDER. 

WONDER foundation

Contact us

The legal profession is rapidly adopting AI, but two High Court rulings reveal the dangers of uncritical reliance. Fake citations generated by AI tools led to regulatory referrals, with judges warning of severe sanctions. AI can aid legal work, but only when lawyers retain control, judgement and oversight.

AI tools promise efficiency in legal work, but recent court cases expose serious risks of unverified or over-reliant use.

<p>Artificial intelligence tools are becoming more deeply embedded in the legal sector, with an increasing number of legal professionals using large language models (LLMs) to help with summarising documents, cross-referencing cases and conducting research. However, recent High Court judgments have laid bare the risks and pitfalls of misusing and over-relying on AI, with the Civil Justice Council setting up an AI working group off the back of the use of several incidents of fake citations submitted to the court during proceedings.</p>
<h2>A cautionary tale from the High Court</h2>
<p>Two recent judgments – <a href="https://www.judiciary.uk/judgments/ayinde-v-london-borough-of-haringey-and-al-haroun-v-qatar-national-bank/" target="_blank" rel="noopener"><em>Ayinde v London Borough of Haringey</em> and <em>Al-Haroun v Qatar National Bank</em></a> – which were referred to the High Court under the ‘Hamid jurisdiction’ (the process by which potential misconduct by legal representatives is addressed), revealed the dangers of mishandling AI tools for case preparation and legal guidance. In both cases, generative AI was apparently used to generate legal citations, some of which were entirely made up. The submissions included invented judgments, fabricated case references and even a non-existent Court of Appeal authority.</p>
<p>In <em>Ayinde</em>, pupil barrister Sarah Forey, instructed by Haringey Law Centre, cited five false cases in judicial review proceedings. Her shifting explanations – ranging from ‘minor citation errors’ to dragging files from a personal box – were eventually dismissed by the judge as implausible. The court found that Forey had either ‘deliberately included fake citations’ or had used ‘generative artificial intelligence tools to produce her list of cases’. Either way, ‘the threshold for initiating contempt proceedings’ had been met.</p>
<p>Describing the incident as a ‘professional shame’ that had tarnished an otherwise strong submission, the judge said the court would not initiate contempt proceedings against Forey, as she was an ‘extremely junior lawyer&#8217;. However, it stressed that its decision should not be interpreted as precedent – Forey’s conduct would still be subject to an investigation by the bar regulator, and ‘lawyers who do not comply with their professional obligations in this respect risk severe sanction.’</p>
<p>In the <em>Al-Haroun</em> case, solicitor Abid Hussain admitted to relying on AI-generated research conducted by his client without verification. The court described this as a ‘lamentable failure’ of professional responsibility and referred him to the SRA, reiterating that lawyers cannot outsource or delegate their professional duty to check sources to clients – or to machines.</p>
<h2>Limits of AI</h2>
<p>These judgments come off the back of numerous reports of AI hallucinating legal citations, with a<a href="https://hai.stanford.edu/news/hallucinating-law-legal-mistakes-large-language-models-are-pervasive" target="_blank" rel="noopener"> 2024 study by Stanford</a> revealing that general-purpose chatbots hallucinated between 58% and 82% of the time on legal queries. With a <a href="https://www.thomsonreuters.com/content/dam/ewp-m/documents/thomsonreuters/en/pdf/reports/future-of-professionals-report-2024.pdf" target="_blank" rel="noopener">growing number of lawyers using AI</a>, there is an increasing need for the legal profession to understand the limits of AI.</p>
<p>In her judgment in <em>Ayinde</em>, Dame Victoria Sharp, President of the King’s Bench Division, said: &#8216;There are serious implications for the administration of justice and public confidence in the justice system if artificial intelligence is misused.&#8217;</p>
<p>The court underlined that lawyers have a professional duty to verify the outputs of AI tools before submitting them to the court or relying on them in advice to clients. Failure to comply could result in sanctions for professional misconduct.</p>
<p>This duty was underscored by Lord Justice Birss. Speaking at London International Disputes Week, he suggested the following advice: ‘You should be taking personal responsibility for what goes in your name, and that applies whether you’re a judge or you’re a lawyer. Looked at that way round, lawyers producing documents with hallucinated case references [would not be] a problem. You shouldn’t be putting anything to a court that you’re not prepared to put your name to.’</p>
<p>Expanding on the example of using AI to summarise documents, Birss said: ‘You can use AI [to summarise], but you can only use it if you have read the document. What you can’t do, is not read the document, and then get AI to summarise it. That’s crazy.’</p>
<h2>The emergence of hallucinations in legal submissions</h2>
<p>An issue addressed by Birss and underscoring both these cases is the problem of ‘hallucinations’ – AI’s tendency to generate plausible but entirely fictitious information. Unlike traditional legal research tools that retrieve content from curated databases, generative AI models (such as GPT-based systems) generate text by pattern recognition, not by verifying against an authoritative source.</p>
<p>In legal contexts, hallucinations can have particularly dangerous consequences. They can lead to misleading citations, improper judicial submissions and ultimately a loss of trust in both the legal profession and the court process.</p>
<p>Leading figures in the AI community are working to resolve the problem of hallucinations, with CEO of Open AI Sam Altman promising a reduced tendency to hallucinate with its anticipated <a href="https://botpress.com/blog/everything-you-should-know-about-gpt-5">ChatGPT 5</a>. However, reports indicate that more advanced models appear to actually hallucinate <a href="https://www.zdnet.com/article/openais-most-capable-models-hallucinate-more-than-earlier-ones/">more</a>, not less.</p>
<h2>Regulators sound the alarm</h2>
<p>The judgments have prompted direct action from the judiciary. The Civil Justice Council (CJC) has now established a <a href="https://www.lawgazette.co.uk/fake-citations-civil-justice-council-sets-up-ai-working-group/5123520.article">working group</a> to examine whether the use of AI in court document preparation requires amendments to the procedural rules. The group is expected to publish its terms of reference shortly and will consider whether new Practice Directions or ethical guidance are needed.</p>
<p>The Bar Council and the Law Society have also been called upon to urgently review their AI guidance and ensure lawyers understand the limitations of AI tools. Ian Jeffery, Chief Executive of the Law Society, acknowledged the seriousness of the issue:</p>
<p>‘This High Court judgment lays bare the dangers of using AI in legal work. We need to ensure the risks are properly addressed.’</p>
<h2>Key risks for legal professionals</h2>
<p>The cases highlight several pressing risks for lawyers and law firms adopting AI:</p>
<ul>
<li>Fake or hallucinated citations: AI-generated references must be verified against trusted databases. Failing to do so can result in professional misconduct findings.</li>
<li>Delegation to clients or machines: lawyers cannot delegate responsibility for research accuracy to clients – or to AI tools.</li>
<li>Damage to reputation: even if courts stop short of contempt, reputational damage and regulatory scrutiny can have lasting consequences on a legal career.</li>
<li>Inadequate training and governance: many legal professionals remain unclear about how AI works and how to use it responsibly. This is continuing to be addressed with efforts concerning <a href="https://www.linnearlegal.co.uk/resources/ai-literacy-making-sure-business-complies-new-eu-regulations/">AI literacy</a>, but more still needs to be done.</li>
<li>Ethical blind spots: overreliance on AI risks eroding the human judgement, care and diligence that define the legal profession.</li>
</ul>
<h2>Striking the right balance</h2>
<p>AI tools offer enormous potential to improve legal service delivery, increase access to justice and boost efficiency – particularly for routine drafting, document analysis and summarisation. But they are only as safe as the people using them. Just as a scalpel can heal or harm, the same AI that speeds up legal work can also mislead courts if misused.</p>
<p>The legal profession must therefore strike a careful balance: embrace innovation, but with robust safeguards. This includes:</p>
<ul>
<li>Implementing AI verification protocols internally.</li>
<li>Training staff on AI limitations and ethical considerations.</li>
<li>Using specialised legal AI tools with curated, trustworthy datasets.</li>
<li>Ensuring all outputs are read, understood and verified by a human lawyer before use.</li>
</ul>
<h2>Don’t skip professional advice</h2>
<p>AI isn’t going anywhere anytime soon, but neither should professionalism. Legal AI should enhance the lawyer’s craft, not replace it. When used responsibly, AI can be a transformative tool. When misused, it risks professional disgrace, sanctions and damage to the justice system itself. Without proper oversight, AI misuse can result in professional misconduct, wasted costs and damage to the integrity of the justice system.</p>
<p>In the longer term, the successful integration of AI in legal practice will depend not on how advanced the technology becomes, but on how well the profession adapts its ethical frameworks, training and culture to ensure it is used wisely.</p>
<p>For more information on AI trends in the legal sector, take a look at our series on <a href="https://www.linnearlegal.co.uk/news-insights/tag/ai-technology/">AI and technology</a>.</p>


AI tools must be used responsibly. Lawyers face professional and reputational consequences for failing to verify outputs or delegating critical tasks to machines.

Risks and pitfalls of using AI in the legal profession

High Court judgments reveal the risks of misusing AI in legal work. Lawyers should verify AI outputs to avoid misconduct or sanctions.

Risks and pitfalls of using AI in the legal profession | Linnear Legal

https://linnearcms.wpenginepowered.com/wp-content/uploads/2025/07/shutterstock_1045047094.jpg

Linnear Legal

https://linnearcms.wpenginepowered.com/wp-content/uploads/2025/01/Linnear-Legal-logo.png

https://secure.gravatar.com/avatar/430378ba79491bc6fd9321b5f889b7c8c3b4952c79a934ef05647cb5b4385d93?s=96&d=mm&r=g

Risks and pitfalls of using AI in the legal profession

Related Insights