On a Wednesday afternoon a trainee lawyer asks her supervising partner whether she may summarise a sixty-page judgment in ChatGPT. The judgment is public, but she wants to paste in the notes from the client meeting so that the summary addresses the case. The firm's AI policy says "carefully" and "no confidential information". The partner reads it, hesitates and tells her she had better not. The trainee does it that evening at home, on her own account.
That is where an AI policy succeeds or fails. A usable policy for a law firm gives clear answers to a handful of questions: which tools, with which account, which client information is off limits, what to do when in doubt, who assesses and how the lawyer sees the decision while typing. Our guide on how to write an AI policy covers the eight questions every AI policy has to answer. This piece is about what changes when the reader is a lawyer.
Why is "no confidential information" not enough?
Because at a law firm almost everything is confidential. The duty of confidentiality covers everything a client has entrusted to the lawyer in that capacity: the facts, the documents, the strategy and often the mere fact that someone is a client. A rule that keeps all of it out of the tool keeps the tool out of the work. And then you get what the trainee did: the work moves to an account the firm cannot see.
A policy that works says what can be done and how. Summarising a public judgment is fine. Pasting in the meeting notes may be fine too, but without the names of the client and the other side, and in the firm's business account. That is an instruction someone can follow on a Wednesday afternoon, and it keeps the work within the environment the firm has a contract for.
Which questions does the policy answer for a firm?
| Question | What a usable answer looks like at a law firm |
|---|---|
| Which tools are allowed? | A list of names with the account next to each: the firm's business ChatGPT licence is a different tool from ChatGPT on a personal account |
| For which work? | For each tool, the tasks that fit (public case law, a clause without parties) and those that do not (a complete file, a client meeting) |
| Which information is off limits? | Who the client is where that is itself confidential, the strategy and settlement range, the other side's documents received under confidentiality, special category data from criminal and family matters |
| What about a tool with no decision? | Public information only, and this is where you request the assessment |
| Who assesses, and how quickly? | The partner responsible for quality or compliance, with IT and the DPO where present, within two weeks |
| Who is covered? | Everyone working under a lawyer's responsibility, so secretarial staff, paralegals and trainees too |
| How do people see a change? | Not in an email, but when they open the tool |
| How does the firm know the policy still holds? | By comparing the list in the policy with the tools that are actually being opened, in numbers, without names |
The last three rows are missing from most policies, yet they are the ones that decide whether the policy still has anything to say, six months on, about the tools in use by then.
Why does the account type come first?
Because what looks like the same tool on screen can be two different services. With a free or personal ChatGPT account, input can be used by default to improve the model unless the user switches that off; with the business versions it is excluded by default, as OpenAI explains. With a free account there is no data processing agreement, and nobody at the firm can manage members, settings or shared links. Our comparison of free and business AI accounts sets out the differences by vendor. A policy that says "ChatGPT is allowed" without naming the account is allowing the free account. So be precise: the business licence is allowed, the personal account is not.
One caveat belongs here. The business account settles the GDPR side: training, administration and the data processing agreement that Article 28 of the GDPR requires once a third party processes personal data for you. It does not settle the duty of confidentiality. The information still goes to a party outside the lawyer's circle, and whether that is acceptable in a particular matter remains for the lawyer to judge. That is why the table has a separate row for what never goes in, even with a business account.
Which client information never goes in?
The general list is in our article on what not to share with AI. A law firm adds three categories: the identity of the client where the fact of the relationship is itself confidential, the strategy and the settlement range, and the other side's documents obtained under confidentiality. What can go in is usually more than people think: a public judgment, a statutory provision, your own clause with [claimant] and [defendant] in place of the names. The rule of thumb the firm writes down: if someone could recognise the matter from the input, it comes out before it goes in.
What happens to a tool nobody has decided on?
A firm of twenty-five lawyers has a policy with three allowed tools. For a few months now, the international practice has been using a translation tool for correspondence with a foreign client. The policy says nothing about it, so nobody asked, and the correspondence names both parties.
The policy then gets a fourth category: no decision yet. Every tool not on the list falls into it automatically, with an instruction everyone can follow: public information only, and report the tool to the quality partner. The translation tool is assessed within two weeks and set to limited use, with a data processing agreement and without party names. Record the vendor's answers with a date, because AI services change their terms faster than a firm revises its policy. A separate article explains why "No decision yet" is the most honest of the four statuses.
How does the policy reach the lawyer while they are typing?
The policy is in the quality handbook, the chatbot is in the browser and the trainee at the start of this piece needed an answer while the text box was open. A policy that exists only in the handbook arrives too late.
The decision works when it appears where the tool is opened. For an allowed tool, nothing happens. For a tool without a decision, a calm notice appears: public information only, with a button to continue. For a tool that is not allowed, the approved alternative is the first button, and continuing takes a deliberate press-and-hold, unless the firm has configured it so that staff cannot continue to that tool. For a limited-use tool, the firm's own condition appears in a single sentence.
BeeSensible does this in the browser: the firm records a decision for each tool, and that decision appears when someone opens the tool. The extension does not block anything of its own accord and does not decide on the firm's behalf; whether someone can still go on to a tool that is not allowed after seeing the notice is a setting the firm chooses, and what happens next is counted without a user ID. When someone pastes client information into a supported AI tool, party names, matter numbers, amounts and company registration numbers are highlighted as they type. The page about AI tools explains how that works. This is the overview, with the decisions shown next to the observed use:
