Vendor AI library · checked November 7, 2026

What AI tools say about the data you put in them.

For each tool: what the vendor's own pages say about training on your content, how long it is kept and the controls you have, on personal and business plans. We quote, link and date every page. Whether a tool fits your business is your decision.

• OpenAI

  ChatGPT

  On personal plans, OpenAI says conversations help train its models unless you turn that off. On its business plans, its privacy page says it does not train on business data.

  Checked November 7, 2026

• Anthropic

  Claude

  On Free, Pro and Max, each user chooses whether chats may be used for model training, and that choice changes how long chats are kept. Team, Enterprise and the API are under separate commercial terms.

  Checked November 7, 2026

• Google

  Google Gemini

  With a personal Google account, Gemini activity can be reviewed by people and used to train models unless Keep Activity is off. In Google Workspace, Google says Gemini does not use customer data for training without permission.

  Checked November 7, 2026

• Microsoft

  Microsoft Copilot

  For organizations, Microsoft says Copilot prompts and responses under enterprise data protection are not used to train foundation models and follow the organization’s retention and audit settings.

  Checked November 7, 2026

• Grammarly

  Grammarly

  Grammarly says its "Product Improvement and Training" setting is on by default for individuals. For business accounts it depends on how the account was bought.

  Checked November 7, 2026

• Notion

  Notion AI

  Notion says neither it nor its AI providers train models on customer data by default, and that its contracts with AI providers prohibit them from training on that data.

  Checked November 7, 2026

• Zoom

  Zoom AI Companion

  Zoom says it does not use audio, video, chat, screen sharing, attachments or other communications-like content to train its own or third-party AI models.

  Checked November 7, 2026

• ElevenLabs

  ElevenLabs

  ElevenLabs’ privacy policy (updated Oct 9, 2026) says it uses personal information to train its AI models, with an opt-out in account settings that applies from then on.

  Checked November 7, 2026

• Otter.ai

  Otter.ai

  Otter.ai’s privacy policy (effective June 16, 2026) says it trains its AI on de-identified audio recordings and transcriptions. The policy does not describe a separate training opt-out.

  Checked November 7, 2026

• Perplexity

  Perplexity

  Perplexity’s privacy notice (updated Oct 23, 2026) lists improving its AI models as a use of data. It does not cover its Enterprise and API offerings, which have separate terms.

  Checked November 7, 2026

Policies change; check the linked pages before you decide. Keep your decision for each tool in an AI register, with the plan you use and who checked the settings. List your AI tools, free. For the laws behind these questions, see the AI and data rules log.