Home › Answers › How do you stop employees pasting client data into ChatGPT?

How do you stop employees pasting client data into ChatGPT?

Why bans alone do not work, and a practical plan a small business can put in place this week.

Short answer

Banning AI rarely works on its own, because people move to their personal phones and accounts, where you can see even less. What works is a short written policy plus an approved tool that is just as easy to use. For client data, that tool should be either a business plan with the right terms or a private AI inside your office, where there is no outside service to leak to.

If your staff are using AI, they are doing it because it saves them time. That is a good sign. The risk is that the fastest way to get a contract summarized is to paste the contract into a personal chat account.

Why doesn't a ban work?

You can block a website on the office network, but you cannot block a phone on a mobile connection. A ban with no alternative pushes the habit out of sight, and you lose what little visibility you had. Employers generally cannot see what someone types into a personal account.

What should you do instead?

  1. Say what is off limits. Name it plainly: client names, financial records, health information, contracts, passwords.
  2. Write one page. A policy nobody reads protects nobody. One page, with a named person to ask.
  3. Give them an approved tool. This is the step that matters. If the approved way is as easy as the unapproved way, most people will use it.
  4. Show people how. Twenty minutes with real examples beats a document.
  5. Review it. Check in every few months as tools and staff change.

Which approved tool should you choose?

 Business cloud planPrivate AI in your office
Where data goesTo the vendor, under business termsNowhere, it stays on your machine
Used for trainingExcluded by default on business plans, check current termsNo, no outside company receives it
Cost modelMonthly, per personOne-time, no per-seat fees
Best fitGeneral work, less sensitive dataClient, patient, and confidential work

A business plan is a real improvement over personal accounts, and for many offices it is enough. If your clients expect confidentiality, or your industry has rules about it, the simplest answer is an AI that never sends the data anywhere. A PrivateOffice Box gives everyone a login, so there is no reason to reach for a personal account.

What if it has already happened?

Assume it has, and do not start with blame. You will hear more if people are not afraid to tell you. Find out what kind of data was involved, and if it could involve client or regulated information, get proper advice on your obligations.

Related: is ChatGPT safe for business data? and how to use AI without exposing client data.

Frequently asked questions

Can I see what employees paste into ChatGPT?

Not on their personal accounts. Business plans offer admin controls, but anything typed into a personal account or personal phone is out of your view.

Does ChatGPT train on business data?

OpenAI says its business plans are excluded from training by default, while consumer accounts may be used unless the person opts out. Check OpenAI's current terms before relying on this.

Should I block ChatGPT on the office network?

You can, but on its own it mostly moves usage to personal phones. It works better alongside an approved tool.

What should an AI policy for a small business include?

What data must never be entered, which tools are approved, who to ask, and what to do after a mistake. One page is enough.

Keep reading