Your Chats Might Be Training Data. Here’s the Switch.

claude series 04 claude chats training data switch Digital Marketing Guru
Share this article:

Key takeaways

  • On Claude’s consumer plans (Free, Pro and Max), a setting called “Help improve our AI models” decides whether your new chats can be used to train future models.
  • Anthropic says data is kept for five years if you allow training, and for 30 days if you don’t.
  • Team, Enterprise and the API are under commercial terms and aren’t covered by this setting. For customer and financial data, that difference matters.

Many business owners paste customer lists, sales figures and supplier contracts into an AI chat without a second thought. Few have checked one setting that decides what happens to those chats afterwards.

The setting

In 2025 Anthropic updated its consumer terms. Users of the Free, Pro and Max plans now choose whether their chats can be used to train future Claude models. The choice is a single toggle:

  1. Open Settings (click your name, then Settings).
  2. Go to Privacy.
  3. Find “Help improve our AI models” and set it the way you want.

On mobile, the path is the same: your name, Settings, Privacy.

What the choice changes

  • On: your new chats and coding sessions can be used for model training, and Anthropic says the data is kept for up to five years.
  • Off: new chats aren’t used for future training, and the standard 30-day retention applies.

Two details people miss:

  • Turning it off is not retroactive for training that has already started. Anthropic says data already used in a training run, or in a model already trained, stays there.
  • Conversations flagged by Anthropic’s safety systems can still be used to improve safety systems, whatever the setting.

Business plans work differently

The consumer setting doesn’t apply to services under Anthropic’s Commercial Terms: Team and Enterprise plans, the API (including through Amazon Bedrock and Google Cloud Vertex AI), and Claude for Government and Education. If your staff use Claude for company work, a Team plan gives you that commercial footing plus admin control over who can use what.

What we recommend to clients

  1. Check the setting today on every personal account used for work.
  2. Move company work to a business plan once more than one or two people use Claude.
  3. Don’t paste what you don’t need to. Remove customer names, phone numbers and ID numbers before asking for analysis. Claude can work with “Customer A, District 3, 5 orders” just as well.
  4. Remember the law. Customer details are personal data under Vietnam’s Personal Data Protection Law, in force since 1 January 2026. Treat AI tools like any other place you store customer data. (This is general guidance, not legal advice.)

Frequently asked questions

Does Claude train on my chats?

On consumer plans (Free, Pro, Max), only if the “Help improve our AI models” setting is on. You can change it at any time in Settings → Privacy. Team, Enterprise and API use are under commercial terms and aren’t covered by this setting.

How long does Anthropic keep my Claude chats?

Anthropic says five years if you allow training, and 30 days if you don’t, for consumer plans.

Is it safe to put customer data into Claude?

Minimise it. Remove names, phone numbers and ID numbers where you can, use a business plan for company work, and treat AI tools like any other system that stores personal data.

Next step: See where your marketing really stands with DMG’s Full Business Audit.

Want help setting up Claude for your team? Talk to DMG.

Claude: What Nobody Tells You, part 4 of 10. Facts checked against Anthropic’s consumer terms update and privacy settings help page on 13 November 2026; settings and plans may change.

Follow DMG on Google: add dmg.guru as a preferred source

Share this article:

Leave a Reply