For years, the legal conversation around AI has focused on the risks of using it. Lawyers have been warned about AI hallucinations, confidentiality breaches and the dangers of relying on inaccurate outputs. But a recent legal statement from the UK Jurisdiction Taskforce (UKJT) has turned that discussion on its head.
In one of the most significant developments yet for the legal profession, the UKJT has concluded that lawyers could, in some circumstances, be negligent not because they used AI, but because they failed to use it when a reasonably competent lawyer would have done so.
The statement marks a significant shift in how AI is viewed within legal practice. AI is not just an optional productivity tool. It is becoming something that could form part of the standard of reasonable professional care.
A shift in professional negligence?
The UKJT’s Legal Statement on Liability for AI Harms concludes that existing English law is already capable of dealing with most disputes arising from AI without the need for AI-specific legislation.
That finding alone is significant. While many jurisdictions continue to debate how AI liability should be regulated, the UKJT argues that established legal principles around negligence, contract and product liability already provide a coherent framework for deciding where responsibility lies.
But the statement’s most eye-catching conclusion concerns professional negligence.
It makes clear that the same legal principles that could make a lawyer liable for using AI carelessly could also expose them to claims for failing to use AI where it would have been reasonable to do so.
Whether that duty exists will depend on what a reasonably competent lawyer in the same field and circumstances would have done. In other words, AI is being treated like any other professional tool. If competent practitioners are expected to use it for a particular task, choosing not to could eventually fall below the required standard of care.
AI as part of legal competence?
Traditionally, competence has meant keeping up with changes in the law and applying appropriate legal knowledge and judgment. The UKJT’s statement suggests that technological competence may now form part of that expectation.
The report gives a practical example. A solicitor handling large-scale disclosure in the Business and Property Courts could potentially be criticised for failing to advise a client about using AI-assisted document review where such tools would improve efficiency and accuracy.
The issue is not that lawyers must always use AI. Rather, they must be able to justify when AI is appropriate and when it is not.
That changes the question firms should be asking. Instead of asking whether they should adopt AI, firms increasingly need to ask whether they can justify not adopting it for particular tasks.
Using AI badly remains risky
The statement is equally clear that using AI irresponsibly creates its own legal risks.
Lawyers remain responsible for every piece of advice they give and every document they submit, regardless of whether AI assisted in producing it.
Professionals could be negligent if they fail to carry out due diligence before selecting an AI system, fail to understand how it works, fail to explain its use to clients where appropriate, or fail to verify AI-generated outputs for errors, bias or hallucinations.
Confidentiality also remains paramount. Uploading privileged or confidential client information into an insecure AI platform could itself amount to a breach of professional duty.
AI does not reduce professional responsibility. It increases the need for careful governance and human oversight.
A changing standard of care
Perhaps the most important implication is that the professional standard of care is no longer fixed.
Historically, the benchmark has always evolved alongside developments in professional practice. The UKJT’s statement suggests AI is following the same path as electronic disclosure, online legal research and other technologies that were once novel but are now routine.
As AI tools become more accurate, more widely adopted and more embedded within legal workflows, expectations of what constitutes reasonable professional practice are also likely to evolve.
What may be considered optional today could become expected tomorrow.
For law firms, that raises important governance questions. Decisions about whether to use AI should be deliberate, documented and based on client interests rather than habit or reluctance to embrace new technology.
Is existing law ready for AI?
The statement also delivers an important message beyond the legal profession. The UKJT concludes that English common law is sufficiently flexible to resolve most AI-related disputes using existing legal principles. Businesses cannot simply blame an AI system for mistakes, nor can organisations avoid liability because a chatbot generated false information.
Instead, responsibility generally remains with the person or organisation deploying the technology.
There are still difficult legal questions, particularly around proving causation in highly autonomous AI systems, and the statement acknowledges that some “liability gaps” may emerge in future cases. However, the report argues these challenges are more practical than conceptual and that English courts have historically adapted successfully to new technologies.
The future is not AI versus lawyers
The legal profession has spent the past two years debating whether AI is safe enough to use. The UKJT’s statement reframes the debate.
The question is no longer whether lawyers should use AI. It may now be whether a competent lawyer could reasonably justify not using it.
That does not mean every legal task should be automated, nor does it diminish the importance of professional judgment. Human expertise remains central to legal practice, especially where strategy, ethics and nuanced legal analysis are concerned.
But the role of AI has fundamentally changed. It is becoming another professional tool that lawyers are expected to understand, govern and deploy appropriately.
For law firms, the implications are big. AI governance is no longer just about managing the risks of adoption. It is also about recognising that, in some situations, failing to embrace AI where it genuinely benefits clients could itself become a professional risk.
The future standard of legal competence may not be defined by whether lawyers use AI, but by whether they know when they should.
