Language: English
Start 14-day free trial

Guide

Claude Team and ChatGPT Business: is your team's data used for training?

Updated

No, by default: Anthropic doesn't use inputs or outputs from Claude Team or Enterprise to train its models, and OpenAI's ChatGPT Business page says the workspace's data is never used for training. The exceptions sit at the edges: on Claude, a chat someone sends as thumbs up or down feedback is kept for up to 5 years and may be used, which an Owner can switch off, and personal plans on both services train on content until each person opts out. For people outside the team plan, such as contractors on their own accounts, keep client data out and share only your method, as skills through AI Skill Pod, which reach their Claude, Cursor or ChatGPT (Pro, Business, Enterprise) from one invite link.

Short answer

  • Claude Team: no training by default.
  • ChatGPT Business: never used for training.
  • The exception: Claude feedback ratings; an Owner can turn them off.
  • Personal plans: on until each person opts out.

Team plans and personal plans, side by side

 Claude TeamChatGPT BusinessPersonal plans
Chats used for trainingNo, by defaultNoYes, until the person opts out
Feedback ratingsKept up to 5 years and may be used; an Owner can turn off Rate chatsNot stated on the Business pageOn Claude, kept up to 5 years, as on Team
Custom data retentionEnterprise onlyEnterprise lists retention controlsNot on Claude's personal plans
Single sign-onYesYesNot on Claude's personal plans

From Anthropic's Privacy Center and Claude's pricing page, read 5 November 2026, 20:31 JST, and OpenAI's ChatGPT Business and pricing pages, read 23:07 JST. Personal plans: Claude Free, Pro, Max and Ultra, and ChatGPT's personal plans.

What an owner checks on day one

  1. Read the terms the plan runs on. Claude Team runs on Anthropic's Commercial Terms, which include its data processing addendum; check ChatGPT Business's terms too.
  2. Decide on feedback ratings. On Claude Team, Organization settings > Data and Privacy has Rate chats; turn it off if no conversation should leave as feedback.
  3. Move work out of personal accounts. Invite the people who handle client data to the team plan.
  4. Write down what stays out of every AI, such as client identifiers you never paste, in the team's AI rules.

A worked example: the firm that asked for a "don't train" switch

Illustration: a made-up business. A nine-person accounting firm moves to Claude Team, and staff ask the owner where to turn training off. The owner:

  • Confirms there's nothing to switch for training on Team, and turns off Rate chats so no client conversation goes out as feedback.
  • Tells the two part-timers on their own Claude Pro accounts to keep client data out, and to opt out of model training in their own privacy settings.
  • Shares the firm's procedures as skills in its pod, so the part-timers' AI follows the firm's method without any client file in it.

AI Skill Pod is this site's product. Pod pricing: Free up to 3 members and 10 skills; Team $5 per seat a month from 3 seats.

Skills for this

All skills

Questions

Does Claude Team train on your data?

Not by default. Anthropic says it won't use inputs or outputs from its commercial products, Claude Team and Enterprise included, to train its models. The exception is a conversation someone sends as thumbs up or down feedback, which can be used, and an Owner can switch that rating off for the organization.

Does ChatGPT Business use your data for training?

No. OpenAI's ChatGPT Business page says the workspace's data is never used for training its models. The personal plans are different: their content is used for training unless each person opts out.

Do we need a 'don't train' setting on a team plan?

Not for training itself, since both team plans leave it off by default. What remains is the people outside the plan: staff or contractors on personal accounts, where training is on until they opt out, so keep client data out of those accounts.

AI skill

Have your AI check a data processing agreement

Put this skill in AI Skill Pod and everyone on your team gets it in their Claude, Cursor or ChatGPT (Pro, Business, Enterprise): one invite link, free to start. Data processing agreement checklist has your AI go through a vendor's data processing terms before the team signs up: who controls the data, subprocessors, security, breach notice, transfers, deletion and audits, with every point for counsel marked.

Use this skill in your AI

14 days free · No card needed

Sources

  1. Anthropic Privacy Center: Is my data used for model training? (commercial) — read at 20:31 JST: by default Anthropic won't use inputs or outputs from commercial products such as Claude for Work to train its models; a conversation sent as thumbs up or down feedback is stored for up to 5 years and may be used for training; a Primary Owner or Owner can switch off Rate chats under Organization settings > Data and Privacy (read 2026-11-05)
  2. Anthropic Privacy Center: How long do you store my organization's data? — read at 20:31 JST: Claude for Work keeps chats so you can continue them; a deleted chat leaves the history at once and the back end within 30 days (read 2026-11-05)
  3. Anthropic Privacy Center: How do I view and sign your DPA? — read at 20:31 JST: the DPA with Standard Contractual Clauses is automatically incorporated into the Commercial Terms of Service (read 2026-11-05)
  4. Claude pricing — read at 20:31 JST: model training is opt-out on Free, Pro, Max and Ultra and none by default on Team and Enterprise; custom data retention controls are Enterprise only (read 2026-11-05)
  5. ChatGPT Business — read at 23:07 JST: "Your data is never used for training our models"; SSO and multi-factor authentication; SCIM, role-based access and data residency come with Enterprise (read 2026-11-05)
  6. ChatGPT pricing — read at 23:07 JST: on the personal plans, content is used to train OpenAI's models, with an opt-out available (read 2026-11-05)