SRA Issues Stark Warning on AI in Legal Work
The Solicitors Regulation Authority (SRA) has put the profession on notice. On August 17, it published a detailed warning notice flagging two specific dangers: AI hallucinations and the leakage of confidential client data through public AI tools.
“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,” the regulator stated. The message is blunt: appropriate human oversight and a risk-based approach are not optional extras.
What the SRA Has Observed
The regulator says it has seen real incidents, not hypotheticals. Solicitors have self-reported AI hallucinations in legal work and court submissions. There have also been reports from senior judges about potential breaches of the Code of Conduct.
Then there’s the data side. Confidential client information has been typed into public AI systems, raising serious data protection and confidentiality concerns. The consequences? Poor client outcomes, slower case progression, and a hit to public trust in the profession.
Accountability Stays With the Solicitor
The SRA takes an outcomes-based approach to regulation. It sets the standards but doesn’t dictate the exact methods. Still, the warning notice makes one thing clear: solicitors and regulated individuals remain accountable for AI output.
Firms need to put in place “effective governance structures, systems and controls” to manage AI risks. That’s not just a suggestion; it’s a compliance expectation.
Court Submissions and Contempt Risk
One of the most serious warnings relates to court submissions. If a solicitor puts AI-hallucinated “false material” before a court, it could be considered contempt of court. The SRA warns of “serious consequences” for those who mislead the courts.
Supervisors of junior or non-authorized colleagues could also be held responsible if false citations end up in court documents. The message is clear: verify every case law authority before submitting it.
Client Data and AI Tools: The Red Lines
Putting client information into a public AI tool will likely breach client confidentiality. The SRA notes that both free-to-use and paid-for AI systems pose risks. Even paid tools aren’t automatically safe.
Client information should only be entered into AI systems where “appropriate contractual, technical and organizational safeguards are in place.” Client data must always remain within a secure environment.
This is a practical concern for many firms. Lawyers often use AI for drafting or research, but the data they input can end up in training models or be exposed to third parties. The SRA is drawing a hard line: confidentiality trumps convenience.
What the Profession Should Do Now
The SRA’s warning notice includes a lengthy list of “considerations” for the profession. Here’s a quick rundown of the key points:
- Ensure solicitors remain accountable for all AI-generated output.
- Implement governance structures to manage AI risks.
- Verify all case law citations are genuine and verifiable.
- Avoid entering client data into public AI tools unless safeguards are in place.
- Keep client data within secure environments at all times.
Brett Dixon, vice president of the Law Society of England and Wales, acknowledged the pace of change. “Technological innovation in legal processes and service delivery is advancing at speed,” he said. He called on the SRA to continue providing “swift and clear guidance” so solicitors can use AI with confidence.
Broader Context: AI Misuse Across Sectors
This isn’t just a UK issue. Similar concerns have emerged globally. In the US, lawyers have been sanctioned for submitting fake citations generated by AI. The SRA’s warning aligns with a broader regulatory trend: AI is a tool, but professional judgment remains human.
For law firms, the takeaway is straightforward. AI can improve efficiency, but it introduces new risks. Firms must balance innovation with their duties to clients and the courts. The SRA’s warning is a reminder that the rules haven’t changed—only the tools have.
If you’re a solicitor, this is the moment to review your AI policies. Check what tools your firm uses, how data flows through them, and whether your supervision structures are robust enough. The SRA is watching.
