What Does It Mean to Be a Competent Lawyer in the Age of AI?

At a conference of law teachers at Western in June, I took part in discussions among professors and lawyers about whether facility with AI should be a core competence in a Canadian law degree at this point in time. This gave rise to larger questions about whether law profs now have an ethical or pedagogical duty of some kind to familiarize themselves with AI and develop methods of teaching students how to use it effectively. This, in turn, pointed to the even larger question of what it means to be a competent lawyer in the age of AI.

Put another way, if you choose to forgo or even resist what some believe to be the many benefits and efficiencies that AI has to offer (as some at the conference said they’ve chosen to do), are you in some way — as a prof or lawyer — failing in your duty of competence?

Have we crossed a certain threshold where you can no longer be competent as a lawyer, can no longer practice or teach to an appropriate standard, given the way that technology has changed realities on the ground?

In 2019, the Federation of Law Societies of Canada amended its Model Code of Professional Conduct to address technological competence. Law societies across the country, including those in Ontario and BC, have since incorporated the Federation’s guidance into their own frameworks.

Briefly, our codes of professional conduct contain a duty of competence, or a duty to possess and apply relevant knowledge, skills, and other attributes described in various sub-rules. The final of these is commonly worded: “otherwise adapting to changing professional requirements, standards, techniques and practices.”

The Federation’s commentary adds:

To maintain the required level of competence, a lawyer should develop an understanding of, and ability to use, technology relevant to the nature and area of the lawyer’s practice and responsibilities. A lawyer should understand the benefits and risks associated with relevant technology…

And:

The required level of technological competence will depend on whether the use or understanding of technology is necessary to the nature and area of the lawyer’s practice and responsibilities and whether the relevant technology is reasonably available to the lawyer.

So, we have both a duty and a standard to meet. We have a duty to become informed about technologies relevant to our practice area, and even to develop an ability to use them.

And how informed must I be and able to use this technology? Well, it depends.

But wait a minute, that’s just what we all want to know!

Do I, as a family or criminal lawyer, an insurance defence litigator or corporate solicitor have to spend time sorting out how to use Harvey, Protégé, or Copilot and decide whether and when to use them in a given case?

Can I not just attend a brief seminar to get an overview of these tools and casually dismiss them as non-essential to the workflow that has served me well for over 20 years?

As a prof, am I not perfectly justified in saying that some of these tools might be quite useful, but my job is to teach students black-letter law, how to think like a lawyer, and so on — things more fundamental than how to use the technology of the day effectively?

Who decides what is necessary here? Are lawyers and profs (and students, for that matter) not free to decide that AI might be useful in some situations, might result in certain efficiencies, but that despite this, they don’t need AI?

And what about AI’s notorious unreliability? How could a tool like Protégé ever become necessary when its research memos and opinions are wrong so much of the time? And isn’t AI like this more broadly?

For that matter, is any technology — aside from pen and paper — really ‘necessary’ to practice law competently? Couldn’t I do a perfectly competent job without even cracking open my laptop if I really wanted to?

Take email. Is it optional at this point? Could you offer a client competent service as a family or employment lawyer if you refused to use it and insisted on writing letters? Probably not. Why? It’s worth spelling out: because your service would be so slow as to risk a failure to resolve a client’s issue in a timely fashion, which it is your duty to do.

So we can point to cases where recent technology becomes necessary; which is to say, where the tech changes realities on the ground and standards of competence shift accordingly.

Is AI one of these cases? Is AI as fundamental to the practice of law as word processing or email?

Can you render legal service competently, that is, in a reasonably timely, efficient, and effective manner, with no reliance on it at all?

Or is it, like email or word processing, something that creates efficiencies so significant as to change the nature and pace of the practice itself?

My sense: not yet, but soon, in some areas, possibly even most, it will be.

Yes, certain AI tools are unreliable for certain purposes. But there are more or less reliable and effective ways of using AI. And some of the more reliable ways of using AI can lead to huge efficiencies. Examples include large document review in litigation, contract drafting when carefully prompted to produce a first draft, case law research done with a frontier model using an MCP, and generating meeting transcriptions and summaries of audio kept on file as a backup.

The efficiencies AI can provide here are so striking that they may soon render the technology necessary to doing a competent job, in the sense that failing to use it would mean wasting a client’s time and money without justification.

But AI is more powerful than email or word processing. If harnessed effectively, it could lead to a higher level of service, raising the bar across the board. I’m thinking here of workflows that lawyers at the frontiers of experimentation with AI in practice have come up with, ranging from automating intake memos and generating draft agreements or routine correspondence, to more elaborate tools for analyzing litigation strategy.

At this point, there is at least a basis for thinking that professional standards of competence may be shifting. The bar may be rising.

We may be approaching a point at which you can no longer be competent in many areas if some portion of your work on a file — research, opinion or contract-drafting, document review — is not in some way aided by AI.

We may not be there yet. We lack clarity about what is ‘necessary’ when it comes to the use of AI in a given practice area. Law profs might continue to ignore the place of AI in the curriculum or remain indifferent to whether facility with AI is, or should be, a core competence in law school.

But soon, I suspect, our collective thinking about this will change. There is at least a real possibility, if not a likelihood, that in the near future we will all need to use AI in some capacity to practice law competently — to deliver service at a level of timeliness and efficiency that clients will be entitled to expect given the general pace at which practice will have evolved.

Again, we may not be there yet. But it’s clearly on the horizon.