The duty of technological competence in practice
By the ChamberOne team 5 minute read
Competence used to mean knowing the law and doing the work with reasonable skill. In many places it now also implies knowing enough about the technology you use to understand what it can do and where it can fail. That does not mean becoming a software engineer. It means that if you put client information into a tool, rely on its output or tell a client how you work, you should understand the basics of what is happening. This article does not say what any rule requires in your jurisdiction, since that varies. It describes what competence tends to look like in daily practice.
Key points
- Read your regulator's or bar's guidance to see how technological competence applies in your jurisdiction.
- For every tool ask what happens to your data, how it can be wrong and what your fallback is.
- Understand that language models generate plausible text and can produce false citations.
- Keep a short register of tools and a quarterly routine for reviewing guidance and outputs.
- Avoid both blind adoption and blanket refusal; learn the basics and check the work.
What the idea asks of you
Professional bodies in several countries have said, in various forms, that lawyers should keep up with changes in the law and its practice, including the benefits and risks of relevant technology. The wording and legal force differ from place to place. Read the guidance from your regulator or bar to see how it applies to you.
The practical reading is modest. You do not need to master every tool. You need to know what the tools you use do with data, where they are likely to be wrong, and how to check their output. And you need to be able to say so if a client or regulator asks.
Three questions for every tool
A short test applies to anything from a document-review tool to a phone app.
- What happens to the information I put in? Where is it stored, who can see it and is it used for anything else?
- How can this tool be wrong, and how would I notice?
- If it fails, such as an outage or an error in output, what is my fallback?
Understanding AI well enough
Language models generate text that is likely to follow from the prompt. They do not look up facts in the way a database does. That is why they can produce fluent statements that are false, including invented citations. It is also why the same question can give different answers on different days.
A lawyer who understands this will check authorities, treat summaries as drafts and avoid giving the tool tasks that require reliable recall. They will also understand that a tool trained on general text may reflect assumptions from other legal systems.
Confidentiality is the part people forget
Competence includes protecting client information, and AI tools are a new route by which it can leave the firm. Before using a tool on client matter, find out whether input is stored, whether it is used to improve the provider's models, where it is processed and who at the provider can access it. If you cannot find out, do not use the tool on client material.
Avoid personal accounts on consumer tools for client work. They are rarely covered by the terms your firm would accept, and nobody else at the firm knows what has been put into them.
A simple routine for staying competent
Competence is not a one-off. Tools change and so does guidance. A routine that fits into a busy week is better than an ambitious plan nobody follows.
- Once a quarter, read any new guidance from your regulator or bar on technology and record what changed.
- Keep a one-page register of the tools in use, what each is for and who is responsible.
- Try each new tool on a non-client example before using it on live work.
- Review one or two real outputs a month with a colleague, looking for errors the tool made.
- Attend a training session or read a short update each term and note it for your records.
- 1Read new guidance from your regulator or bar
- 2Update the one-page register of tools
- 3Try new tools on a non-client example
- 4Review real outputs with a colleague
- 5Note any training for your records
- Then back to step 1.
Supervision and delegation
Partners are responsible for how the people they supervise use technology. A junior who pastes a client's contract into a public chatbot has created a problem for the partner as well as for themselves. That is a reason to set clear rules, not just to rely on good sense.
Suppose a four-lawyer firm writes a one-page policy: which tools may be used, what may be entered, who to ask if unsure, and that all AI output is reviewed by the lawyer before use. It then spends ten minutes at a team meeting going through two examples. This will not prevent every error. It does mean that when something goes wrong, the firm can show it had communicated its expectations.
Telling clients about your tools
Some clients will ask how you work, and some regulators and clients expect to be told when technology plays a material part. Check what applies to you. Plain, honest answers help: 'We use an AI tool to help prepare first drafts and summaries. A lawyer reviews everything before it reaches you, and your documents are handled under the terms in our policy.' If you cannot give an honest answer in two sentences, you may not yet know enough about your tools.
Documenting what you did
If a client or regulator asks how a piece of work was produced, a short record answers it. Note the tool used, the task, what was entered at a general level, who reviewed the output and what they changed. A line or two on the matter file is enough. Over time the notes also show which tools earn their place and which create more checking than they save.
Do not record more client detail than you need. The record should show the process, not duplicate the work.
Avoiding two opposite mistakes
One mistake is to adopt tools enthusiastically with no checks. The other is to refuse to engage because the risks seem too many. Both can be a failure of competence. A lawyer who never learns what a widely used tool can do may miss ways to serve clients faster or more cheaply, while a lawyer who uses it blindly exposes clients to errors.
The middle position is unglamorous. Learn the basics, try things carefully, check outputs, protect data and keep a record of what you did. That is what a reasonable colleague would expect.
- Tools used enthusiastically without checking
- Exposes clients to errors
- Never learning what a widely used tool can do
- May miss ways to serve clients faster or more cheaply
This article is general information about the law, legal practice and technology, current when it was published. It is not legal advice. Laws and guidance change: check the sources linked above and take advice on your own circumstances from a lawyer qualified where you are.
Read next
Client confidentiality and AI tools: a checklist before you paste anything in
The questions to answer about retention, training use, access, hosting and sub-processors before client information goes into any AI tool.
5 minute readShadow AI in law firms: staff are already using it, so what next?
Why bans fail, how to find out what is being used, and how to offer an approved alternative with clear rules.
5 minute readWriting a law firm AI policy: an outline you can adapt
Six sections that a short, usable AI policy needs, with the questions each one should answer.
4 minute read