Free AI Is All You Need to Supercharge Your Practice

The market for legal AI is teaming with options. Many of them are compelling. All of them are expensive.

What the companies offering these tools hope that you don’t notice is that free (or almost free) tools, like ChatGPT, Claude, and Gemini are getting so good at basic legal research tasks that many lawyers looking to boost their productivity don’t need to look any further.

We highlight three ways to use these free tools to make research more effective — but we note that becoming a subscriber to one of them, for roughly $30 a month, will tend to get you better results.

General overviews

Many lawyers haven’t moved beyond their first impressions of ChatGPT formed in the months following its release in 2022, when much of its output was laughably false or erroneous. It’s come a long way since.

The free AI platforms are now so good at providing an accurate general summary of an area of Canadian law that they now serve as the best first place to start when you’re new to an area.

However, a subscription to one of these tools will give you access to a “reasoning” model, like ChatGPT o1, which breaks a query into steps, building further verification into the process.

We ran this prompt into the various models: “Summarize the law on unjust enrichment in Canada in 500 words, with citations to leading case law or relevant statutes.”

The best result came from ChatGPT o1, which you can read here.

It cites three Supreme Court of Canada decisions, including the foundational authority, Pettkus v Becker. Citations to all three cases were correct, and the summary — though basic and general — is accurate.

We were also impressed to see that the answer noted the limited relevance of provincial legislation on point, correctly citing both BC and Ontario family law statutes.

Running the query on the various free versions of the frontier models rendered largely the same content with one or two citations inaccurate or made up. But showing results to law professors who work on unjust enrichment confirmed that the general overviews were correct.

None of this would suffice for a detailed memo on point — and by no means should you trust the citations! But for someone simply seeking to find their bearings in a new area of law, these tools could be invaluable.

Boolean searches

The various frontier models have recently unveiled a new tool similar to OpenAI’s Deep Research that will soon make it possible to describe a fact pattern and ask the AI to visit a site like Canlii to search for cases. It will be able to read the results and sift through them for you. We’re not there yet. But we’re likely no more than a year away.

In the meantime, one helpful use of a free tool like ChatGPT or Claude is to ask it to assist in formulating a Boolean search string, which you would plug into Canlii or another case law database.

Many of us have honed this skill over time. But for both beginners and more seasoned researchers, the main utility of using AI for this lies not in formulating the string, but in speeding up the process of finding the right one.

For example, we asked Claude to: “Craft a boolean search string for a search on a Canadian legal database of cases on the law of unjust enrichment involving partners in a restaurant business who had no formal contract and in which the court applies the test in Pettkus.”

The result:

“(“unjust enrichment” OR “unjustified enrichment”) AND (restaurant* OR cafe* OR dining OR bistro*) AND (partner* OR co-owner* OR proprietor*) AND (“no contract” OR “absence of contract” OR “without contract” OR “informal agreement” OR “informal arrangement”) AND (“Pettkus v Becker” OR Pettkus) AND (enrich* /p deprivat* /p correspond*)”

Plugging this into Canlii brought up a dozen cases, including three directly on point.

You may not get the right search string off the bat, as often happens when we do this on our own. Using AI might involve a few rounds of trial and error, providing feedback to the program as you run the queries. But it will likely be quicker than doing this on your own and give you new ideas.

Case summaries

Our last example is using free AI to do case summaries. Here too the aim is not to avoid reading cases or to rely on AI entirely.

But to get a quick and accessible overview of a case, AI can do magic. We ran this query into the free version of ChatGPT: “Summarize Moore v. Sweet, 2018 SCC 52 in 300 words” — and you can read the summary here.

Comparing it with the case itself, all the essential details are here, and they’re correct: the factual matrix, the issue, the test applied, and the outcome.

This could come in handy when coming across a case cited in a factum or in another case — to get a quick sense of what it’s about. Or it might help you decide what cases you turn up on Canlii are worth delving into.

These are only three possible uses of free AI tools for legal research. There are no doubt more. And the technology is only becoming better by the day.

Free AI tools help level the playing field, making it possible for any lawyer to become more efficient in practice.