Ever wonder what your lawyer really thinks about you using AI?
More of our clients are leaning on ChatGPT (and tools like it) during their separation. We get it – it’s free and available at 11pm when you can’t sleep. But we are also seeing the cost of it. Matters become more complex, more expensive, and harder for us to actually run.
Here are the traps.
You can lose privilege
Legal professional privilege protects the strategic conversations between you and us. If you paste our advice into ChatGPT to “translate it” or get a second opinion, you may lose that protection. AI tools are third parties. Once privileged advice has been shared with a third party, the other side may be entitled to ask for it.
If something we have said is confusing, just ask us. We will explain it again.
Don’t ask AI to decode the other side’s letters
Those AI conversations are discoverable. If your matter ends up in court, the other side may be entitled to see them – including your speculation about their motives and the rude things you said about your ex at midnight. That is not a position we want you in.
Reading between the lines of opposing counsel’s letters is our job. Send them to us.
Tell us what you actually think – don’t polish it through AI
Relationship property work is largely judgment. We are listening for what you really want, what worries you, and where you might give a little. If you run your messages through AI first, all that signal gets sanded off.
It also goes the other way. AI-polished emails using legal phrasing that doesn’t quite match what you mean leave us guessing. We end up needing more emails, calls and meetings with you. That costs you money, time and stress.
A blunt, rambling email from you (exactly like what you would plug into AI to correct) is more useful than a polished AI version. We can work with the real one. We have been interpreting people in your shoes for years.
AI gets things wrong – particularly New Zealand law
A few things to know… even if you tell AI to act like a New Zealand relationship property lawyer, it can:
- often draw on overseas law. But the rules in United States and United Kingdom are very different to the rules in New Zealand.
- invent cases. Confidently, with realistic-looking citations. Lawyers overseas have been disciplined for citing hallucinated cases that did not exist. People representing themselves have now been told off twice by our Supreme Court this year.
- write to please you, to give you the answer it thinks you want. Your lawyer will give you the answer you need.
If AI tells you something that surprises you, ring us before you act on it or lose sleep over it.
Your chat history is not as private as it feels
There are many criminal cases where someone’s Google search history has been used as evidence of their state of mind. ChatGPT history is no different in principle. It is rare for our matters to go to court, but if yours does, those conversations can paint you in an unhelpful light. Be careful what you type.
Where to from here?
Use whatever helps you feel prepared. Just remember that prepared isn’t the same as advised – and your instincts about your own situation are probably better than you think. The AI-polished version of your question often strips out the thing we actually needed to hear. Talk to us in your own words, messy and uncertain. That’s genuinely more useful.
When things shift in your matter, when you have a decision to make, when something the other side does surprises you – that’s when you call us. Your matter involves real law, a real person on the other side, and real consequences. AI is working from patterns. We’re working from your file, and we usually know the other side’s lawyer. This area of law rewards judgment and patience more than information, and that particular combination isn’t something you can Google.
We’re on your side. That’s not something that changes based on how you got here or what you searched at 2am. When you’re ready to move, we’re here. Book a free call.
Comments (0)