Chicago Business Attorney Blog
The uncomfortable truth: your AI conversations may be evidence
If you are using AI tools like ChatGPT, Claude, or Google Gemini to ask legal questions, draft contracts, or think through business decisions, you need to understand one thing:
Those conversations are likely not protected by attorney-client privilege.
And that means they may be discoverable in litigation.
Not hypothetically. Not someday.
Now.
Why AI is not your lawyer (even when it feels like one)
Attorney-client privilege exists for a very specific reason. It protects confidential communications between a client and a licensed attorney acting in a legal capacity.
AI tools do not meet that standard.
Even if:
•You are asking legal questions
•The answers sound authoritative
•The output resembles legal advice
There is no attorney. No legal duty. No privilege.
From a legal standpoint, using AI for advice is closer to:
•asking a colleague
•searching Google
•or drafting notes to yourself
And none of those are protected.
The real risk: discoverability
Here is where this becomes a business problem, not just a legal technicality.
If your company is involved in litigation, opposing counsel can request:
•Internal communications
•Decision-making records
•Drafts and revisions
•Digital tool usage
That can include:
•AI chat logs
•prompts entered by employees
•outputs relied on in business decisions
If those conversations contain:
•legal strategy
•risk discussions
•admissions or uncertainty
You may have just handed the other side a roadmap.
A scenario most companies are already walking into
An employee uses AI to ask:
“Can we terminate this employee without legal risk?”
They paste details. Names. Facts.
The AI responds with a confident answer.
That conversation is saved.
Months later, the terminated employee files a claim.
During discovery, that AI exchange becomes:
•a timestamped record
•a statement of intent
•potential evidence of knowledge or disregard
And it is not privileged.
The illusion of confidentiality
Many businesses assume:
•“It’s private.”
•“It’s internal.”
•“It’s just a tool.”
But most AI platforms:
•store interactions
•may use data for training or improvement
•are governed by terms of service, not legal privilege
Even enterprise versions require careful review of:
•data retention
•access controls
•contractual protections
This is not just a tech issue. It is a legal exposure issue.
Why this matters more in 2026 than it did a year ago
AI is no longer experimental inside companies.
It is embedded in:
•marketing teams
•HR workflows
•operations
•executive decision-making
At the same time:
•courts are catching up
•regulators are paying attention
•litigators are getting smarter about where to look
AI usage is creating a new category of evidence.
Most companies have not adjusted.
What businesses should do now
This is fixable, but only if you act intentionally.
1. Create an AI usage policy immediately
Define:
•what employees can and cannot input
•prohibited topics (legal, HR, confidential data)
•approved tools
2. Separate legal advice from AI experimentation
AI can support:
•efficiency
•drafting
•brainstorming
But it should not replace:
•legal analysis
•risk evaluation
•privileged communication
3. Train your team like this is a real risk, because it is
Most exposure comes from:
•well-meaning employees
•convenience
•speed
Not bad intent.
4. Review your contracts and vendors
If you are using AI tools:
•understand their data policies
•negotiate terms where possible
•ensure alignment with your risk tolerance
5. Involve legal early, not after the fact
If AI is part of your operations, your legal strategy needs to reflect that.
The bottom line
AI is powerful. It is efficient. It is not confidential.
If you are treating AI like a lawyer, you are creating risk you cannot see yet.
But opposing counsel will.
A better way to think about it
Use AI as a tool.
Use your lawyer for advice.
Know the difference.
About the author
George Bellas is a business attorney advising companies on corporate law, risk management, and emerging legal issues including artificial intelligence, compliance, and governance.
If your employees are using AI, you already have legal exposure
Most companies do not realize where the risk is until it shows up in a lawsuit, a demand letter, or discovery.
By then, it is too late to fix.
George Bellas works with companies to identify exactly where AI is creating liability and how to shut it down before it becomes a problem.
•Where are you exposed right now?
•What are your employees putting into AI tools?
•What could be discoverable tomorrow?
If you cannot answer those questions with certainty, you have a gap.
Do not wait for opposing counsel to find it first.
Contact George Bellas today to assess your risk, implement an AI policy, and protect your business before it costs you.