Ravi: British law can overcome difficulties
Lawyers and other professionals could be sued for negligence for failing to use artificial intelligence (AI), according to the latest legal statement from the UK Jurisdictional Taskforce (UKJT).
The UKTJ also cited examples where professionals could be held liable for a lack of “reasonable care and skill” in the use of AI, including failing to carry out appropriate due diligence, failing to explain to clients how the AI was being used, and failing to check output for errors or bias.
UKTJ said its previous legal pronouncements on crypto assets and smart contracts have been “directly adopted by UK judges” and relied upon by international investors who have chosen UK law for their technology transactions.
It concluded that existing UK law already provides the legal framework needed to resolve most AI liability disputes and no new legislation is needed.
among them Legal statement on liability for AI harmthe task force stated: “Whether there is a breach of duty in not using the AI will, of course, be determined by whether a reasonable expert of comparable rank/specialty should have used the AI in that context.
“In that case, again, reliance would be placed on professional body regulation and guidance (if any) and expert evidence regarding the conduct of competent professionals in the field.
“The potential for breach of duty arising from failure to use AI reflects the fact that AI in the hands of professionals is a tool. The questions of whether such tools should be used, and if so, how, are no different from the questions that arise with respect to other tools available to professionals.”
An example of where a solicitor could be held liable for failing to use AI is cited by the UKTJ as “where a business and property court solicitor fails to advise a client that he or she may wish to consider some form of AI-assisted tool to review a large number of documents”.
Looking at instances in which professionals are likely to be held liable for lack of due care and skill when using AI, the task force cited examples of “failure to conduct appropriate due diligence on an AI system before using it in a client’s operations, particularly if the system is new, innovative, or comes from an untested provider.”
Experts must be able to “explain (at least in broad strokes)” to their clients how the AI they are trying to use works.
“Putting a client’s confidential or privileged information into an AI system that is not properly secured and confidential is very likely a breach of duty to the client,” the lawyer said.
Experts must also “ensure that sufficient testing has been carried out to ensure that the AI system is adequate and appropriate for the task for which it is envisaged.”
“Professionals who fail to supervise the output of AI systems are likely to be deemed negligent,” and oversight “may require professionals to identify and address risks of error or bias in AI systems.”
The UKTJ said that professionals can be held liable for information generated by chatbots, and that liability is “generally established where a legal entity claims that an AI chatbot is communicating on its behalf, or where there is an express or implied representation that what the chatbot says is correct.”
However, while Canadian courts have held airlines liable for false statements made by chatbots, there is currently no UK authority on “the question of whether AI can speak ‘on behalf of’ a legal entity.”
Matthew Ravi KC, Chairman of the UKTJ Drafting Committee, commented: “We have reached the view that a new liability regime is not needed to provide remedies for AI harm and that UK law’s existing approach to liability already provides a consistent framework for analyzing liability.”
“While there may be difficulties in some factual scenarios, UK law has long had the ability to meet the challenges of new technology and we believe it can do so with respect to AI as well.”
Sir Geoffrey Vos, Master of the Rolls, said in his foreword to the legal statement: “Since the advent of generative AI in 2022, lawyers and AI users alike have expressed concerns about when and if developers of large-scale language models and those involved in the AI supply chain will be held accountable for harm caused by AI.
“This legal statement seeks to provide much-needed legal certainty and legal predictability in an area where few cases have yet come to court.”
He added: “At a time when the capabilities and uses of generative AI are rapidly evolving, we believe this legal statement will be of immense value to the legal community, the technology community, and users.”
