Why Law Firms Shouldn't Use ChatGPT for Client Documents

If you're a lawyer and you've pasted a contract, a client email, or case notes into ChatGPT — you're not alone. Millions of professionals do it every day. But for law firms specifically, the privacy implications deserve serious attention before it becomes a disciplinary issue.

What actually happens when you use ChatGPT

When you send text to ChatGPT, that content is transmitted to and processed on OpenAI's servers. By default, OpenAI's terms allow them to use your conversations to improve their models. While there is an opt-out setting for business accounts, the reality is:

• The data leaves your control the moment you hit enter. It travels over the internet to servers you don't own.
• Client information is in that data. Names, case details, deal terms, medical histories — whatever you pasted in.
• You can't verify what happens to it after the conversation ends.

For most businesses this is an acceptable tradeoff. For a law firm, it isn't.

The professional conduct problem

Bar associations across the US have begun issuing ethics guidance on AI use. The core concern is consistent: lawyers have a duty of confidentiality under Model Rule 1.6, and that duty extends to how information is handled by technology tools.

Sending privileged client information to a third-party cloud service — even a sophisticated one like OpenAI — without client consent raises real questions. Several state bars have already flagged this explicitly, and enforcement guidance is still being written. Getting ahead of it now is far easier than responding to a complaint later.

What a private AI installation actually looks like

The alternative isn't going without AI — it's running AI that never leaves your building. Tools like Ollama combined with Open WebUI give you a full ChatGPT-style interface that runs on your own computer or server. Your questions and documents never leave your network.

From a practical standpoint:

• Document summarisation — paste a 100-page contract and get a plain-English summary. Nothing goes to any external server.
• Drafting assistance — ask it to draft a letter or clause. The output stays local.
• Research notes — ask it to explain case law or statute language you've pasted in.
• Staff use — multiple fee earners can access it through their browser on your local network.

The experience is comparable to ChatGPT. The risk profile is completely different.

The cost comparison

A private AI setup from Dexal LLC runs $150–$1,500 as a one-time fee depending on complexity. Compare that to:

• ChatGPT Team: $25/user/month
• Microsoft Copilot for Microsoft 365: $30/user/month
• Ongoing exposure under your professional obligations

For a firm of even five people, the private setup pays for itself within a few months — with the compliance peace of mind included.

---

If you want to understand whether your current hardware can support a private AI setup, book a free compatibility check. I'll give you a straight answer with no obligation.