At half past five a corporate lawyer has a thirty-page purchase agreement in front of her and a client who wants to know what it says by nine tomorrow morning. She pastes the text into a chatbot and has a summary within a minute. In the corridor a colleague mentions a contract analysis tool he has been trying out since a conference. Nobody did anything wrong, yet in both cases client information went to a vendor the firm has no agreement with.
That is what AI governance at a law firm is about. Not whether lawyers may use AI, because they already do, but which tools are in use, what the firm has decided about each of them, whether the lawyer knows that decision while the file is sitting in the text box and whether the firm can show afterwards that its rules were followed. Our article on what AI governance is covers the subject in general. This piece is about what a law firm adds to it: the duty of confidentiality.
What do lawyers do with AI?
Three things, mostly. Summarising: a file, a stack of exhibits, a forty-page judgment. Drafting: the first version of a summons, a letter to the other side, a clause that needs tightening. And research: finding case law, exploring a legal question, testing an argument.
Around that sits a quieter layer. The firm's legal database now has an AI feature, the dictation software produces summaries, a transcription tool writes up client meetings and a translation tool helps with a cross-border matter. All of them process text from case files, and most have never been decided on because nobody saw them as AI tools. Our article on shadow AI explains how that layer builds up.
| Task | What the input usually contains | The question for the firm |
|---|---|---|
| Summarising a file or exhibits | Party names, matter number, facts, positions | May this go into this tool, with this account? |
| Drafting a summons or agreement | Parties, amounts, the client's strategy | Does the strategy stay within the circle of confidentiality? |
| Searching case law | Often just a legal question, sometimes the facts of the matter | Do the facts give away who the client is? |
| Writing up a meeting | Everything the client said | Has any tool been assessed for this at all? |
Why does confidentiality carry more weight than the GDPR?
For most organisations the question of AI and data is a GDPR question: personal data, legal basis, data processing agreement. A law firm has a further duty on top of that, one that is older and reaches further. A Dutch lawyer (advocaat) has a statutory duty of confidentiality covering everything entrusted to them in that capacity, and a matching right to refuse to give evidence about it. That right exists for the client, not the lawyer: a client must be able to speak freely.
Two things follow. The duty covers more than personal data: a client's negotiating strategy may contain no name at all and still fall under it. A data processing agreement settles the GDPR side, as our article on data processing agreements for AI tools explains, and leaves the professional duty untouched. And the duty generally extends to everyone working under the lawyer's responsibility: the secretarial staff, the paralegal, the trainee who writes the summary.
In the Netherlands, lawyers are supervised by the dekens (the local bar presidents), and the Netherlands Bar sets the professional rules. What it means when information from a file ends up in an AI tool depends on the tool, the account and the situation. What applies in general is that the firm must be able to explain the judgement it made, and "we did not know that tool was being used" is not a judgement. Following reports of data breaches through AI chatbots, the Dutch Data Protection Authority advises organisations to set out which tools are allowed and which data may go into them. You can only do that for tools you know exist.
What does this look like at a firm of forty lawyers?
A firm has bought a business ChatGPT licence and allowed the AI feature of its legal database. The policy runs to two pages and the partners have adopted it. Then, for the first time, the firm looks at which AI services are actually being opened from the browser.
Alongside the two allowed tools, Claude and Perplexity turn out to be in use, as well as a transcription tool for client meetings, the contract analysis tool from the conference and a translation tool for a cross-border matter. Three of the five are not recorded anywhere. The lawyer using the transcription tool had no reason to think it was not allowed.
This is not an incident. A policy describes what the firm knew on the day it was adopted, and use has moved on since. The real work starts here: choosing a status for each tool. The transcription tool is "Not allowed" until there is a data processing agreement and it is clear how recordings are stored. The translation tool is set to "Limited use", without party names. The contract analysis tool keeps the status "No decision yet", and until a decision is made only public model contracts go into it. The four statuses are explained in a separate article.
BeeSensible links a catalogue of over 1,000 assessed AI tools to the use observed at the firm: only the domain of an opened tool, at most once a day per device, with no user ID and no page content. An administrator sees in the Inbox which tools have no decision yet and records the decision. Here is the overview:
