The regulator for the UK’s legal sector has reminded solicitors and law firms of their obligations to use AI safely and responsibly, after highlighting two key areas of concern.

The Solicitors Regulation Authority (SRA) published a lengthy warning notice on August 17 sounding the alarm over AI-related hallucinations and data leaks.

“We are concerned that some of those operating in the regulated legal sector are not fulfilling their obligations to uphold their duties to the courts, clients and third parties, and are not always following all relevant rules and regulations,” it noted.

“Appropriate human oversight, informed professional judgement and a proportionate, risk-based approach are essential to ensure compliance with your regulatory and legal obligations.”

Read more on AI misuse: Manchester Law Firm Leads 15,000 to Sue Google and Microsoft over AI Data

The SRA said that it has observed:

• AI hallucinations in legal work and court submissions – including both solicitors self-reporting and reports of potential breaches of its Code of Conduct from senior members of the judiciary
• Cases of confidential client information being entered into public AI tools, raising data protection and confidentiality concerns

Such incidents mean clients get poor outcomes, case progression slows, and public trust in the profession is damaged, the SRA warned.

Advice for the Legal Profession

The SRA takes an outcomes-based approach to regulation, meaning it sets the standards it expects of the sector but does not prescribe how those standards should be met.

Its warning notice featured a lengthy list of “considerations” for the profession, including:

• Solicitors/regulated individuals remain accountable for AI output
• Firms must put in place “effective governance structures, systems and controls” to manage AI risks
• Solicitors who mislead the courts risk “serious consequences” and must ensure that all submissions of named case law authorities are genuine, relevant and have a verifiable citation
• Putting AI hallucinated “false material” before the court could be considered contempt of court
• Those who supervise “junior or non-authorized colleagues” could be held responsible if they put false citations before the court
• Putting client information into a public AI tool will likely breach client confidentiality, with both free-to-use and paid-for AI systems posing risks
• Client information should only be entered into AI systems where “appropriate contractual, technical and organizational safeguards are in place to protect confidentiality”
• Client data must always remain within a secure environment

Law Society of England and Wales vice president, Brett Dixon, added that solicitors have a duty to use AI and other technologies responsibly in order to act in the best interest of their clients.

“Technological innovation in legal processes and service delivery is advancing at speed,” he added. “As the use of AI evolves rapidly, the SRA will need to continue providing the profession with swift and clear guidance so solicitors can use AI with confidence in line with their duties.”