Could You Be Negligent for Not Using AI?

The UK Jurisdiction Taskforce published its latest legal statement on July 9. In it, the taskforce made clear that lawyers and other professionals can face negligence claims for failing to use artificial intelligence.

This is not speculation. It is a formal legal position from a body whose previous statements on cryptoassets and smart contracts have been adopted directly by English judges and relied upon by international investors choosing English law for technology transactions.

The negligence threshold

Whether a professional has been negligent for not using AI depends on whether a reasonable professional of comparable rank or specialism should have used it in that context.

The UKJT’s statement is specific:

Whether there is a breach of duty in failing to use AI will, of course, be judged according to whether a reasonable professional of a comparable rank/specialism should have used AI in that context.

The standard is the same as it always has been for professional tools. A solicitor is not negligent for not using Word because there are perfectly good alternatives. But if a reasonable solicitor would use Word in a given situation, the failure to do so becomes relevant.

The UKJT gives one concrete example: a solicitor in the Business and Property Courts failing to advise a client that it may wish to consider some form of AI-assisted tool to review large volumes of documents.

The implication is straightforward. If a reasonable solicitor would use AI to handle document review in a particular case, the failure to even consider it may constitute a breach of duty.

The flip side: using AI poorly

The negligence risk is not just for failing to use AI. Using it carelessly is also a breach of duty — and arguably the sharper risk, since firms rushing to adopt AI under compliance pressure may cut corners exactly where it matters.

The taskforce identifies several scenarios where professionals are likely to be found negligent:

• Failing to conduct proper due diligence on an AI system before using it for client work — particularly if the system is new, innovative, or from an untested provider
• Failing to explain to clients how the AI works, at minimum in broad terms
• Failing to ensure the AI system has been sufficiently tested for suitability to the task
• Failing to exercise oversight of the AI system’s outputs, including identifying and addressing the risk of errors and biases

These are not abstract concerns. They map directly to the kinds of failures already surfacing in practice. A US federal court recently stripped legal privilege from documents created in a public AI tool, and the Mexican government lost 195 million taxpayer records after a hacker spent a month exploiting an AI system. Both are cautionary tales about what happens when oversight fails.

The statement also flags a specific security risk: a solicitor putting confidential or privileged information into an AI system that is not suitably secure and confidential is highly likely to be in breach of duty.

The security question

This last point is the one most likely to keep firms awake at night.

The UKJT’s position is clear: professionals must understand not just how AI works, but where the data goes and who can access it. If a firm is using a cloud AI service, the question is not whether the provider has security measures in place — it is whether those measures are adequate for the data being processed and whether the firm has independently verified them.

This is the problem cloud AI cannot solve for law firms. No matter how robust a provider’s security claims, the fundamental issue remains: client data leaves the firm’s control. As we detailed in our analysis of why UK high street firms face an impossible choice with cloud AI, the professional conduct obligations around confidentiality create a structural tension that vendor assurances alone cannot resolve.

The answer that more firms are reaching for is capability sovereignty — retaining full control over both the AI tools and the data they process, by running models on infrastructure the firm owns and controls. Our earlier piece on capability sovereignty explores why this principle matters beyond marketing and into the heart of professional duty.

Matthew Lavy KC, chair of the UKJT’s drafting committee, put it this way:

We have reached the view that providing remedies for AI harms does not require a new liability regime, and that English law’s existing approach to legal liability already provides a coherent framework for analysing where responsibility should fall.

The point is that existing law covers AI liability. No new legislation is needed. The duty to exercise reasonable care and skill applies to AI in the same way it applies to any professional tool.

What this means for firms

For the legal profession, the UKJT’s statement is a signal. AI is no longer an optional add-on. It is a tool that reasonable professionals are expected to use in appropriate situations — and the expectations extend well beyond the courtroom. For in-house teams in particular, AI can be the difference between a stretched practice and a genuinely effective one, as we explored in our earlier analysis.

But that comes with obligations. Firms need to:

• Assess which tasks benefit from AI and which do not
• Select systems that are appropriate for the work — including understanding the provider’s security posture and data handling practices
• Train staff to use AI effectively and to understand its limitations, not just its capabilities
• Maintain oversight of outputs, not just in theory but through documented processes

The firms that treat this seriously will be better positioned to demonstrate that they have met their professional obligations. The firms that ignore it may find themselves answering questions about why they did not act when a reasonable professional would have.

The practical path

The UKJT’s statement does not tell firms to buy into every AI product on the market. It tells them to make considered decisions.

That means understanding what AI can and cannot do, what security risks it introduces, and what controls are appropriate for different types of work. It means having policies, not just opinions, about when and how AI is used.

For many firms, this is a question of building competence rather than purchasing capability. The technology is widely available. The real challenge is deploying it responsibly — with proper due diligence, adequate training, and genuine oversight.

If you are working through how on-premises AI fits into this framework, our complete guide to on-premises AI in the UK walks through the practical considerations from hardware to governance.

Getting this wrong is a risk. Not using it at all, in appropriate circumstances, may be one too.

---

JD Fortress AI deploys secure, on-premises AI for law firms across the UK. If you are exploring whether AI is right for your practice and how to get started, get in touch for a confidential discussion — no pitch, just practical talk.