In August, the Information Commissioner’s Office (ICO) published its draft corporate strategy for 2026–2028, narrowing its focus to four main priorities. One of the four was aimed squarely at AI: promoting trust and transparency in AI.
For a regulator that could have chosen almost anything, that feels like a signal. Not just about the impact AI is already having on everyday business, but a bigger shift in what clients expect. AI transparency in law firms, being open with clients about when and how AI is used on their matter, how their data is handled, and who stays accountable for the result, is fast becoming a question of client trust, not just compliance.
The strategy is transparent about the gap it’s trying to close. Around 20 million UK adults now use AI. This is almsot double the number from last year, and yet, in the ICO’s own words, trust lags behind. With only 33% of people in Great Britain saying they trust companies that use AI to protect their personal data, adoption is clearly running ahead of confidence.
As a sector built on trust and confidentiality, the UK legal profession is likely to feel this apprehension most deeply. This is why the ICO’s message hits harder here than almost anywhere else. Get AI transparency in your law firm right, and it becomes something your clients welcome. Get it wrong, and it quietly erodes the very relationships your firm depends on.
Clients aren’t against AI, they just want to be told
It would be easy to assume the lack of trust is simply public resistance, that clients are afraid of change. Yet, the evidence suggests otherwise. The Legal Services Board’s June 2026 research found that most people are actually optimistic about AI in legal services. 70% expect it to make legal help easier to access, 66% more accessible, and 64% more affordable.
But that optimism comes with conditions. Consumers drew a set of firm red lines, five protections they would not trade away for the convenience of AI:
- A minimum guarantee of accuracy
- No consequential action without informed consent
- Some level of human oversight
- Access to redress if things go wrong
- Protection of their personal data
Two of these depend entirely on one thing: being told.
A client cannot give informed consent to something they were never told was happening, and human oversight only reassures a client who knows it is there. Seen this way, transparency isn’t a courtesy laid on top of AI use, it’s what makes the use acceptable in the first place.
Most law firms can’t yet answer for their AI
Adoption is now mainstream; it’s no longer something firms are debating. LexisNexis found that 84% of UK lawyers are now using generative AI or planning to, a standing start two years ago, to mainstream implementation today.
But adoption has run ahead of accountability, and the client has been left out. In a 2026 survey, 68% of in-house legal professionals said they had no idea whether their external law firms were using AI at all. The technology has arrived; the conversation with the client largely hasn’t.
The gap isn’t the adoption. It’s that firms have moved fast without deciding how they’d answer a question a client could reasonably ask tomorrow: is AI being used on my matter, and how? Being transparent about something you can’t account for is impossible. It makes AI transparency a governance question before it’s communications one.
Transparency is becoming a selection criteria for clients
This is no longer only about reassurance; it’s starting to shape who wins the work. 82% of general counsel say they expect their law firms to clearly explain how AI is used in their work.
The same shift shows up in the terms clients set. As we noted in our guide to client confidentiality and AI, a growing number of clients (banks, insurers, government bodies) now write their own AI expectations into engagement letters. A clear answer to “what do you use, and what happens to our data” is fast becoming something clients weigh when they decide who to instruct. Silence is no longer neutral; increasingly, it reads as a firm that hasn’t thought about it, and that costs the instruction.
What AI transparency requires from law firms
The regulator has already set the expectation. The SRA’s risk outlook on AI is direct: “Tell clients when you will be using AI with their case, and how it will operate.” Its more recent compliance guidance goes further. It should always be made clear to clients where they are interfacing with AI. And responsibility sits squarely with the firm: “You remain responsible and accountable for the outputs from the AI you are using.”
The cost of getting it wrong is no longer hypothetical, either. The SRA’s 2026 warning notice addresses a number of cases of AI misuse, including fabricated case citations reaching court.
But being open about your AI is only possible if you can account for it. You cannot tell a client what you can’t yourself see. This is why transparency has to be built on governance, not bolted on after the fact.
Three questions your clients could ask about your AI usage
A practical way to test where your firm stands is to imagine a client asking three questions, and to notice whether you have a clear answer to each:
“What AI tools will you use on my matter?”
You can only answer if you know what’s live across the firm, including the AI built into software you already use. Our Pre-Deployment AI Governance Checklist is designed to establish exactly that, tool by tool.
“What happens to my confidential information?”
The answer turns on control, whether a tool is contractually closed to the firm or an open, public one. Our Client Confidentiality Safe-Use Guide works through where that line sits, including the situations where it’s less obvious than it looks.
“Will you tell me when you use it?”
This needs a decided position, not an improvised one. Our AI Use Policy template sets out three ways a firm can handle disclosure; proactively, through engagement terms, or on request, with the trade-offs of each, so the choice is a deliberate one.
If you can answer all three cleanly, you’re already transparent in the way clients and regulators are asking for. If you can’t, the fix isn’t a statement on your website, it’s the governance underneath it.
AI transparency is an opportunity, not a burden for law firms
Done properly, AI transparency in law firms isn’t a brake on adoption. It’s what allows clients to say yes to AI with confidence. As the ICO puts it, “building that trust is essential to unlocking the full economic potential of AI.”
For a law firm, that’s the whole opportunity in a sentence. The firms that can answer for their AI use won’t just avoid the risks, they’ll be the ones clients feel safe bringing their most important work to.
And transparency doesn’t end at disclosure. Once clients can see the AI in your work, the next question is what it means for how you bill for it. That’s something we explore in AI and the billable hour.