Over the past year, we’ve noticed more and more clients mentioning that they’ve asked ChatGPT, or another AI platform, to look over a letter of advice, a settlement offer, or a set of medical reports before coming to see us. It’s an understandable instinct: AI tools are fast, free and available at midnight when a worry won’t let you sleep. But using AI to review legal documents or legal advice can create risks, particularly during a personal injury claim.
Those risks can include confidentiality and privacy concerns, inaccurate legal information and, potentially, issues with legal professional privilege.
Here is why we ask clients to speak with their lawyer before running documents from their claim through ChatGPT or another publicly available AI platform.
Using AI for legal documents isn’t a problem by itself
We use AI tools ourselves, in appropriate and controlled ways, to improve how efficiently we work. This isn’t a blanket objection to AI or new technology.
The concern is what happens when confidential information about your legal matter is entered into an external AI platform.
Information shared between you and your lawyer can have particular legal protections. Once information about your claim leaves that confidential relationship, you may not be able to get those protections back.
Could using ChatGPT affect legal professional privilege?
If you’ve received advice from us about your workers’ compensation claim, your CTP matter or a public liability case, that advice is usually protected by legal professional privilege.
That protection exists because the advice was given, and received, in confidence.
Copying that advice, a settlement offer, an independent medical examination report or even a summary of what we’ve told you into an AI platform can undo that protection. Once it’s been shared outside our relationship, whether or not the AI platform says it’s private, there’s a real risk that protection is gone for good, and that can affect your position if your matter ends up in dispute or before the Tribunal or a Court.
Think carefully before entering confidential information into an AI platform
There is also a straightforward privacy issue.
Information you enter into a publicly available AI platform can be stored, used to train the system, accessed by people who have nothing to do with your case, or held on servers overseas.
For a claim that involves your medical history, your income, your family circumstances or details about other people, that’s a significant amount of sensitive information to hand over to a system you don’t control.
AI can sound convincing and still give incorrect legal advice
One of the biggest risks of relying on AI for legal advice is that an answer can sound clear, confident and authoritative while still being wrong.
We’ve seen AI-generated material contain:
- incorrect calculations of noise exposure history
- misstated limitation periods
- invented case references
- incorrect explanations of legislation
- answers that fail to account for the client’s particular circumstances
For example, an AI-generated explanation may not accurately reflect how the Return to Work Act, the Civil Liability Act or other relevant legislation applies to a particular claim.
AI tools don’t know your specific facts, and they aren’t checking their own work against the law as it stands today.
If incorrect AI-generated information is relied upon, or put before a Court or Tribunal, it can create additional work and potentially affect the credibility, cost and conduct of a matter.
If you’ve already used AI on your matter, tell us
This isn’t about getting anyone in trouble. If you’ve asked an AI platform to review a document, explain some advice, or give you a second opinion on your claim, we would genuinely rather know.
It lets us check the facts, correct anything that’s gone off track and make sure your confidentiality hasn’t been compromised in a way we need to manage.
What should you do instead?
If something in your advice doesn’t make sense, or you want a second look at anything we’ve sent you, call us. That’s what we’re here for.
It means the advice you’re relying on has actually been checked by someone who knows your file, the law that applies to it, and how it fits together.
AI has its place, but your claim deserves advice that’s been properly verified, kept confidential, and tailored to your circumstances, not generated by a system that’s never seen your file
If you have any questions about your claim or about anything in this article, please get in touch with our team.
