AI Regulation Tracker  /  Courts and guidance

BC Provincial Court Urges Verifying AI Before Filing

On July 8, 2026, the Provincial Court of British Columbia published "Guidance on using AI to prepare for court," an informational eNews item that tells both lawyers and self-represented litigants to check AI output before they file. This is voluntary informational guidance, not a binding practice direction and not a sanction, and it changes no legal duty.

By Anthony Guerriero  ·  The Leveraged Years Regulatory Research  ·  Published July 15, 2026  ·  Last updated July 15, 2026

On July 8, 2026, the Provincial Court of British Columbia published a short piece of guidance about using AI to prepare for court. It ran as an eNews item, which is the Court's informational channel, not the place where it issues formal practice directions. That framing matters, and I will come back to it, but the substance is what most people filing in that court will care about.

What the guidance actually says

The core warning is about accuracy. AI tools, and generative chatbots in particular, can invent cases and citations that do not exist, a failure often called a hallucination, and they can state the law incorrectly while sounding perfectly confident. The Court's point is that a filer cannot take AI output on faith. It puts the responsibility on the person filing to check the material first.

The Court states it plainly. Whether you are a lawyer or a litigant, when you file any material with the Court, you need to check it carefully to make sure it is accurate. That is the whole idea in one sentence, and it applies the same way to a self-represented person as it does to counsel.

On the how, the guidance is practical. It recommends independently verifying any case or citation against a reliable source such as CanLII or the court databases rather than trusting that the AI got it right. It suggests hyperlinking the authorities you rely on so the Court can follow them directly. And it says that if a judge asks whether you used AI to prepare your material, you should answer truthfully.

None of that is exotic. It is the same verification discipline a careful professional already applies to any research. What is worth noting is that the Court felt the need to spell it out for everyone who files, not just for the bar.

Why this one is different from what we already track

We already track the professional-conduct side of this. Law societies across Canada and bar regulators in the US have reminded their members that competence and candor duties apply when they use AI, and that filing fabricated citations can draw discipline. Those are professional-conduct bodies speaking to the lawyers they license.

This is a court speaking directly to the people who appear before it, and that includes self-represented litigants who answer to no law society at all. A professional-conduct rule does not reach a self-represented person. A court's expectations about what shows up in a filing do. That is the gap this guidance fills, and it is why it belongs in the tracker as its own item rather than as a footnote to the law-society material.

It is also a reminder of where the exposure sits. A lawyer who files a fabricated case risks a conduct complaint on top of the Court's reaction. A self-represented litigant who does the same still puts a bad filing in front of a judge, still wastes the Court's time, and still has to answer for it, even without a regulator in the picture.

What this is, and what it is not

I want to be precise about the weight of this, because it is easy to see a court's name and assume a new rule landed.

This is voluntary informational guidance. It ran as an eNews item, which is the Court telling people how to prepare, not a practice direction that changes procedure and not a rule that carries a defined penalty. It does not impose a new filing requirement, it does not create a disclosure obligation, and it does not announce a sanction. Nobody's legal duties changed on July 8 because this guidance was posted.

What it does is set expectations and inform behavior. When a court tells everyone who files that it expects verified, accurate material and truthful answers about AI use, that is a clear read on how that court is thinking. The underlying duties it points back to, filing accurate material and being candid with the Court, already exist. The guidance simply makes them explicit in the AI context. Treat it as the Court showing its hand, not as new law.

What this means for anyone filing

If you file in the Provincial Court of British Columbia, the practical instruction is direct. Do not file AI-generated cases or citations you have not confirmed exist and say what you think they say. Check every authority against CanLII or the court databases, link them, and if a judge asks about AI use, answer honestly. That is the whole ask, and it applies to counsel and self-represented litigants alike.

For US lawyers and finance and legal professionals watching from outside Canada, the value is comparative. The Provincial Court of BC is not your court, and this guidance has no effect on any US filing. But it lines up neatly with the wave of AI-verification messages coming out of North American courts, including the standing orders and individual-judge requirements that have appeared across US federal courts asking parties to disclose or certify AI use and to verify citations. The direct US legal effect here is low. The signal is worth reading. Courts on both sides of the border are converging on the same expectation, which is that a human verifies AI output before it reaches the bench.

The common thread is verification. Whether the message comes from a Canadian provincial court, a US district judge's standing order, or a law society, the answer that keeps you out of trouble is the same one: confirm the source exists, confirm it says what you claim, and be honest about how the document was prepared.

What to do now

Read the guidance as what it is, a voluntary reminder to verify, not a new rule or a sanction. If you file in BC Provincial Court, build a habit of checking every AI-surfaced case against CanLII or a court database before it goes in, hyperlink your authorities, and be ready to answer truthfully if asked about AI use. If you are a self-represented litigant, take the same care a lawyer would, because the Court holds your filing to the accuracy standard regardless of who prepared it. And if you practice in the US, treat this as one more data point in a clear cross-border trend, then follow your own court's actual standing orders and rules, which are the ones that bind you.

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal advice. Confirm how any court expectation or rule applies to your situation with qualified professionals in the relevant jurisdiction.