AI in law firms: opportunity, responsibility and risk

This article has been collaboratively written by Lee Mulligan of Howden and Clare Hughes-Williams and Peter Given of DAC Beachcroft. 

Howden and DAC Beachcroft are pleased to share their perspectives on the use of artificial intelligence (AI) in law firms and their suppliers. AI is increasingly being adopted across businesses to drive innovation, improve efficiency and support decision-making. At the same time, its use brings important legal, regulatory and governance considerations that law firms should understand and manage effectively.

This guide provides a high-level overview of some of the key issues and practical considerations law firms may wish to take into account when adopting and using AI. As the technology continues to evolve at pace, so too will the legal and regulatory landscape, as well as market practice and expectations surrounding its use, as the Misuse of AI Warning notice from the SRA in August 2026 evidences. 

One thing is clear: AI is here to stay. The challenge for organisations is not whether to engage with AI, but how to harness its benefits responsibly, effectively and with appropriate oversight. Firms that combine human judgement with AI-enabled insight are likely to outperform firms that rely solely on either. 

How firms are using AI

AI is increasingly being integrated into a range of legal and business processes, helping firms improve efficiency, manage information and support decision-making. 

• Business operations/processes and internal workflows.
• Legal research, including analysing case law and legislation. 
• Predictive analytics, e.g., modelling possible outcomes and identifying patterns from data.
• Data collection and disclosure.
• Document review/analysis to reduce review time and increase consistency.
• Drafting/transcription, producing first drafts of documents or automating document generation.

AI adoption can outpace supervision and governance arrangements and that gap can remain invisible until something goes wrong. If errors arise, firms may be required to evidence not just that controls exist, but that they were effective in practice. 
 

What is not changing?

Lawyers' ongoing professional duties including accountability to their clients, client privilege and confidentiality, supervision requirements and obligations to the Court remain unaltered.  AI is not a substitute for exercising professional judgement.

Professional obligations arise from:

• The contract: the scope of the retainer with the client must be agreed, including warnings about and consent to use AI.
• Lawyers’ obligations under their respective codes of conduct. 
• Common law duties in tort to undertake work using reasonable skill and care. What would a reasonably competent solicitor do?
• Duty to the Court to be truthful and to uphold the proper administration of justice.

See guidance from (but not limited to) the Law Society; the Solicitors Regulation Authority; the Bar Standards Board; the Civil Justice Council and judgments which address the responsible use of AI. 
 

What can go wrong?

While AI can deliver significant benefits, its use also introduces risks that can lead to regulatory, professional and financial consequences if not properly managed.

Examples of circumstances where a professional may be liable for using or failing to use AI in the provision of their services include:  

• Using AI inappropriately.
• Failing to conduct proper due diligence on supply chains. Do AI vendors’ security standards meet the firm’s requirements?
• Failing to use AI when a competent member of the profession would have done so.
• Using AI where it should not be used – e.g. letting AI make judgement calls.
• Using an unsuitable model – e.g. a general purpose tool for a task that requires a specific output.
• Failing to test AI or validate its outputs effectively. Supervision and/or approval may be required – e.g., hallucinations in Court documents.
• Failing to understand the capabilities and limitations of the AI. Technology may sound authoritative whilst being inaccurate. 
• Confidentiality/data issues, including inadvertent waiver of privilege or confidentiality. Do you know how data will be stored or used? 
   

Making this work in practice

Effective AI use depends on clear policies, appropriate supervision and regular review to ensure risks are identified and managed as technology evolves.

Firms should:

• Identify, monitor and manage material risks through comprehensive and accessible policies and procedures and training to ensure staff are competent and understand the risks.
• Talk to your broker - does your PI programme reflect how your firm uses AI?
• Implement effective governance structures to ensure that the firm complies with existing regulatory and legislative requirements through supervision. For example, how are skeleton arguments in which case law is cited being checked?
• Put active, ongoing monitoring of AI policies in place, working with IT colleagues. The pace of change is such that annual reviews will not suffice.  
• Ensure that AI vendors’ security standards meet the firm’s requirements. For example, are they advocating the use of a "closed" system to protect client data?

About our authors

Lee Mulligan, Howden

Since joining Howden to lead risk management for professional services Lee has been building a professional services view of AI risk, helping firms translate governance thinking into practical, renewal-relevant conversations. If you would like to discuss your use of AI and its insurance implications, please contact her.  

Clare Hughes-Williams and Peter Given, DAC Beachcroft

DAC Beachcroft has extensive experience advising on claims, disputes and regulatory issues, including those arising from the use of AI. Please reach out to Clare Hughes-Williams, who is recognised for her expertise in solicitors' regulatory and disciplinary matters, and Peter Given, who advises on all aspects of data protection, privacy law and regulatory investigations. 

Contact details for all three of our authors are available below.

This article has been written in conjunction with Clare Hughes-Williams and Peter Given of DAC Beachcroft. and the opinions and views stated by them in this article are those of Clare Hughes-Williams and Peter Given on behalf of DAC Beachcroft and not Howden Insurance Brokers Limited (“Howden”). Howden is an insurance broker and is not authorised or regulated to advise on AI in law firms. Howden shall not (i) owe or accept any duty, responsibility or liability to you or any other person; and (ii) be liable in respect of any loss, damage or expense caused by your or any other party’s reliance on this article.

DAC Beachcroft publications are created on a general basis for information only and do not constitute legal or other professional advice. No liability is accepted to users or third parties for the use of the contents or any errors or inaccuracies therein. Professional advice should always be obtained before applying the information to particular circumstances. For further details please go to www.dacbeachcroft.com/en/gb/about/legal-notice. Please also read our DAC Beachcroft Group privacy policy at www.dacbeachcroft.com/en/gb/about/privacy-policy. By reading this publication you accept that you have read, understood and agree to the terms of this disclaimer. The copyright in this communication is retained by DAC Beachcroft. © DAC Beachcroft