AI tools and client data

Is Codex safe for client data?

Short answer: Whether Codex is safe for client data turns on which plan your staff are on, which account they signed in with, and what you have told them they may put in it — not on the product itself. On Plus and Pro, OpenAI may use code, instructions and task context to train its models unless somebody turns that off; on Business, OpenAI says it does not train its models on code, instructions and task context, by default; on Enterprise and Edu, OpenAI says it does not train its models on code, instructions and task context, by default. Every fact on this page is OpenAI's own, quoted from its own pages and dated, and where we did not find a statement this page says so rather than filling the gap. Obsidian Ridge's free 30-day AI security assessment shows which AI tools your staff actually use.

By Kfir Yair, CISSP. Published . Updated .

See which AI tools your staff use

Plan by plan, in OpenAI’s own words

Codex is OpenAI's coding agent, available through desktop, command-line, IDE and cloud interfaces.

What you are askingPlus and ProBusinessEnterprise and Edu
Trains the vendor's models by defaultYes, unless training is turned off in ChatGPT data controls. This statement covers Plus and Pro.“Conversations may be used to improve models unless you turn off training in ChatGPT data controls.” source, read No, by default for business data; the Enterprise and Edu plans include the Business commitment.“No training on your business data by default.” source, read No, by default for business data; the Enterprise and Edu plans include the Business commitment.“No training on your business data by default.” source, read
How to turn training offYes, unless training is turned off in ChatGPT data controls. This statement covers Plus and Pro.“Conversations may be used to improve models unless you turn off training in ChatGPT data controls.” source, read Not found in the pages we read.Not found in the pages we read.
How long data is keptNot found in the pages we read.Not found in the pages we read.Not found in the pages we read.
What an admin can see or controlNot found in the pages we read.Not found in the pages we read.Enterprise and Edu include retention and residency controls; confirm the available options for the workspace.“Data retention and data residency controls” source, read
Single sign-onNot found in the pages we read.The business workspace includes SAML single sign-on.“SAML SSO” source, read The business workspace includes SAML single sign-on.“SAML SSO” source, read
Business Associate AgreementNot found in the pages we read.Not found in the pages we read.Not found in the pages we read.
Where the data is processedNot found in the pages we read.Not found in the pages we read.Not found in the pages we read.

An empty cell is not a reassurance and not a criticism. "Not found in the pages we read." means the OpenAI pages listed under Sources, read on the dates shown, do not answer it — so there is nothing for us to repeat, and something for you to ask OpenAI before client information goes near it.

What this means for

The duty does not change because the tool is new. What changes is where the client information ends up.

  • Law firms. A client confidence pasted into a personal account has left the firm’s control. Our law firm page sets out the duties we build evidence for; the decision here is whether a personal account may touch a client matter at all.
  • Dental and medical practices. We didn't find a BAA statement in the pages listed below; ask OpenAI before any patient data goes in. See dental practices.
  • CPA and accounting firms. Your written information security plan has to describe the tools people actually use, not the ones you approved. See accounting firms.
  • Financial advisers. Client account data in a personal assistant is the exposure; see financial advisers.
  • Software teams. A coding assistant reads the repository and can run commands, so the question is what it can reach, not what it was asked. See what AI agents on work laptops can reach.

How to tell which version your staff use

The plan is the whole answer, and you cannot tell from the outside. Three ways to find out, in order of how much they prove.

  1. 1. Ask

    Quick, free, and incomplete. People forget the tool they installed once and the account they signed in with.

  2. 2. Look at the billing and the admin console

    This finds what the business pays for. It does not find a personal account, which is the one that matters here.

  3. 3. Look on the computers

    A tool somebody installed themselves never appears in an IT list. Obsidian Ridge’s free 30-day AI security assessment takes a quiet inventory for 30 days and shows what each tool could reach, then a CISSP goes through it with you. See the assessment, or read a sample set of findings.

Settings to check today

From OpenAI’s own documentation. Each one is a change somebody can make this afternoon.

  • Review the separate Include environments setting in Codex. The ChatGPT training switch and Privacy Portal opt-out do not change it.“Changing your ChatGPT setting or opting out through the Privacy Portal does not change that setting.” source, read

Worth knowing

  • Codex environment context has its own training control; review it as well as the ChatGPT control described in the plan rows.“separate Include environments setting” source, read

Questions we get asked

Is Codex safe for client data?

Whether Codex is safe for client data turns on which plan your staff are on, which account they signed in with, and what you have told them they may put in it — not on the product itself.

On Plus and Pro, OpenAI may use code, instructions and task context to train its models unless somebody turns that off; on Business, OpenAI says it does not train its models on code, instructions and task context, by default; on Enterprise and Edu, OpenAI says it does not train its models on code, instructions and task context, by default.

Every fact on this page is OpenAI's own, quoted from its own pages and dated, and where we did not find a statement this page says so rather than filling the gap.

Does OpenAI train its models on what we put into Codex?

On Plus and Pro, OpenAI may use code, instructions and task context to train its models unless somebody turns that off; on Business, OpenAI says it does not train its models on code, instructions and task context, by default; on Enterprise and Edu, OpenAI says it does not train its models on code, instructions and task context, by default.

On Plus and Pro: Yes, unless training is turned off in ChatGPT data controls. This statement covers Plus and Pro.

OpenAI's own words: "Conversations may be used to improve models unless you turn off training in ChatGPT data controls." (read 2026-11-05).

On Business, Enterprise and Edu: No, by default for business data; the Enterprise and Edu plans include the Business commitment. OpenAI's own words: "No training on your business data by default." (read 2026-11-05).

How long does OpenAI keep what we put into Codex?

On Plus and Pro, Business, Enterprise and Edu: Not found in the pages we read.

Can an administrator read what we put into Codex?

On Plus and Pro, Business: Not found in the pages we read. On Enterprise and Edu: Enterprise and Edu include retention and residency controls; confirm the available options for the workspace. OpenAI's own words: "Data retention and data residency controls" (read 2026-11-05).

Will OpenAI sign a Business Associate Agreement for Codex?

We didn't find a BAA statement in the pages listed below; ask OpenAI before any patient data goes in.

How do we find out which version of Codex our staff are using?

Look on the computers. A tool somebody installed themselves, or signed into with a personal account, does not appear in any IT list. Obsidian Ridge's free 30-day AI security assessment takes that inventory for 30 days and shows what each tool could reach.

Sources

OpenAI’s own pages, with the date each was read. Nothing on this page comes from anywhere else.

All 47 tools, side by side, or run the AI tool risk checker over the ones you know about.

Other tools people ask about

Same question, different answers. The ones in the same category are the ones staff swap between.