Does Claude train on your data?
Anthropic · AI chat assistant · official site ↗
Depends on your planZDR / DPA on enterpriseOpt-out available
Does Claude train its models on your data?
It depends — Claude trains on consumer/free-tier data by default but excludes business, enterprise and API tiers.
Since 28 Aug 2025, Anthropic trains on consumer Claude (Free, Pro, Max) chats and coding sessions unless you opt out, but does not train on Claude for Work/Team/Enterprise/Gov/Edu or the API.
Can you opt out?
Consumer users must opt out in Privacy Settings (a one-time pop-up was shown; existing users had until 8 Oct 2025 to choose). Opting out keeps the original 30-day retention; allowing training extends retention up to 5 years. Commercial and API usage is excluded from training and has nothing to toggle.
Zero retention / DPA
API inputs and outputs are deleted after 7 days (reduced from 30 days on 14 Sep 2025) and are never used for training. Qualifying enterprise customers can sign a Zero Data Retention agreement so prompts/outputs are not stored beyond abuse screening. DPA ↗
What the listicles get wrong
This is a dated reversal: Anthropic previously did not train on consumer chats and deleted within 30 days. ZDR can be voided when Claude is accessed via third-party clouds (e.g. Amazon Bedrock, Google Vertex), which run under those providers' terms.
Verdict by plan tier
Free / Pro / Max (consumer)Trains unless you opt outTrains on chats and coding sessions by default since 28 Aug 2025. Allowing training extends retention to 5 years; opting out keeps 30-day retention.
Claude for Work / Team / Enterprise / Gov / EduNo trainingGoverned by Commercial Terms; not used to train models. The consumer policy change does not apply.
API (Anthropic-hosted)No trainingNever used for training; 7-day retention. ZDR available to qualifying enterprise customers.
Get notified when this changes
We track Claude's data-training and retention policy. Leave your email and we'll send one note if it changes.
Frequently asked questions
Does Claude train its AI models on my data?
It depends — Claude trains on consumer/free-tier data by default but excludes business, enterprise and API tiers. Since 28 Aug 2025, Anthropic trains on consumer Claude (Free, Pro, Max) chats and coding sessions unless you opt out, but does not train on Claude for Work/Team/Enterprise/Gov/Edu or the API.
Can I opt out of Claude training on my data?
Consumer users must opt out in Privacy Settings (a one-time pop-up was shown; existing users had until 8 Oct 2025 to choose). Opting out keeps the original 30-day retention; allowing training extends retention up to 5 years. Commercial and API usage is excluded from training and has nothing to toggle.
Does Claude offer zero data retention (ZDR) or a DPA?
API inputs and outputs are deleted after 7 days (reduced from 30 days on 14 Sep 2025) and are never used for training. Qualifying enterprise customers can sign a Zero Data Retention agreement so prompts/outputs are not stored beyond abuse screening.
Is Claude safe to use with confidential or proprietary data?
It depends on your plan tier. Free / Pro / Max (consumer): Trains on chats and coding sessions by default since 28 Aug 2025. Allowing training extends retention to 5 years; opting out keeps 30-day retention. Claude for Work / Team / Enterprise / Gov / Edu: Governed by Commercial Terms; not used to train models. The consumer policy change does not apply. API (Anthropic-hosted): Never used for training; 7-day retention. ZDR available to qualifying enterprise customers. Always confirm current terms with Anthropic before sending confidential data — this is cited public information, not legal advice.
Sources
https://www.anthropic.com/news/updates-to-our-consumer-terms
https://techcrunch.com/2025/08/28/anthropic-users-face-a-new-choice-opt-out-or-share-your-data-for-ai-training/
https://privacy.claude.com/en/articles/10301952-updates-to-our-privacy-policy
This page is cited public information, not legal or compliance advice. Whether Claude trains on your data, and any zero-retention or DPA option, can depend on your specific plan, region and contract. Always confirm current terms with Anthropic before sending confidential or proprietary data.