Provision record
Anthropic · Claude.ai Terms of Service · View original document ↗

Training Data Use and Opt-Out Carve-Outs

High severity Common · 290 of 352 platforms
Stay ahead of the changes
Track Anthropic and get the diff the day its terms change.
Share 𝕏 Share in Share 🔒 PDF
Document Record

What it is

Anthropic can use your Claude conversations to train its AI by default, and you can turn this off in settings — but even if you do, clicking thumbs up or down on a response, or having any message flagged for safety, means that content still gets used for training.

This analysis describes what Anthropic's agreement states, permits, or reserves. It does not constitute a legal determination about enforceability. Regulatory applicability and practical outcomes may vary by jurisdiction, enforcement context, and individual circumstances. Read our methodology

ConductAtlas Analysis

Why it matters (compliance & governance perspective)

The provision establishes the scope of permitted data uses and creates a conditional opt-out structure rather than unconditional data use restrictions. The carve-outs ensure training data availability for specific operational purposes—feedback incorporation and safety detection—even when users exercise the general opt-out option.

Clause Stability Stable

0
Changes
6
Months Monitored
Apr 27, 2026
First Seen
Apr 27, 2026
Last Seen
This clause type exists across 5149 other provisions on other platforms.

Consumer impact (what this means for users)

If you opt out of model training but ever rate a Claude response or have a conversation flagged for safety review, your conversation data is still used to train Anthropic's AI — limiting the practical effectiveness of the opt-out for most active users.

What you can do

⚠️ These actions may provide transparency or partial mitigation but may not fully address the underlying issue. Effectiveness varies by jurisdiction and individual circumstances.
  • Opt Out of Arbitration
    Log in to Claude.ai, navigate to Account Settings, and locate the model training or data usage preference toggle. Disable the option to use your conversations for model training. Note that this opt-out does not apply to feedback you provide or to content flagged for safety review.

How other platforms handle this

Glassdoor Medium

We will also provide an individual opt-out choice, or opt-in for sensitive data, before we share your data with third parties other than our agents, or before we use it for a purpose other than which it was originally collected.

Square Medium

to request that your data be transferred to a third party (data portability)

Google Cloud Medium

Your organization may allow you to access and export your data in order to back it up or transfer it to a service outside of Google.

See all platforms with this clause type →
▸ View Original Clause Language DOCUMENT RECORD
"
We may use Materials to provide, maintain, and improve the Services and to develop other products and services, including training our models, unless you opt out of training through your account settings. Even if you opt out, we will use Materials for model training when: (1) you provide Feedback to us regarding any Materials, or (2) your Materials are flagged for safety review to improve our ability to detect harmful content, enforce our policies, or advance our safety research.

Excerpt from Anthropic's Claude.ai Terms of Service

ConductAtlas Analysis

Institutional analysis (regulatory & governance intelligence)

(1) REGULATORY FRAMEWORK: This provision implicates GDPR Art.

Insight

Unlock the full institutional analysis

Enforcement risk, jurisdiction flags, contract triggers, and due diligence action items.

Applicable agencies

  • Federal Trade Commission (ftc)
    Oversees unfair or deceptive business practices and can investigate companies that mislead consumers about data collection, sharing, or use.
    Who can file: Anyone affected by the company's practices (US or international)
    What you need: Your account details, a timeline of relevant events, and a description of the specific issue
    What to expect: Complaints inform FTC enforcement priorities and investigations but do not result in individual resolution or compensation
    File a complaint →
  • State Attorney General
    State AGs in California, New York, Texas, and other states can investigate violations of state consumer protection and privacy laws, including CCPA (California), SHIELD Act (New York), and equivalents.
    Who can file: Residents of states with comprehensive privacy laws — primarily California, Virginia, Colorado, Connecticut, and Utah
    What you need: Evidence of the violation, explanation of how your state rights were affected, and your account or contact information with the company
    What to expect: Outcomes vary by state. May result in investigation, enforcement action, or requirement for the company to change practices. No direct individual compensation in most cases.

    Search "[your state] attorney general consumer complaint" to find your state's direct complaint form

Applicable regulations

EU AI Act
European Union
BIPA
Illinois, USA
CCPA/CPRA
California, USA
Colorado AI Act
US-CO
Connecticut Data Privacy Act Amendments
US-CT
CAN-SPAM
United States Federal
EU AI Act - High Risk Provisions
EU
FTC Act Section 5
United States Federal
GDPR
European Union
Indiana Consumer Data Protection Act
US-IN
Kentucky Consumer Data Protection Act
US-KY
UK GDPR
United Kingdom
Universal Opt-Out Mechanism Expansion 2026
US

Provision details

Document information
Document
Claude.ai Terms of Service
Entity
Anthropic
Document last updated
May 5, 2026
Tracking information
First tracked
March 6, 2026
Last verified
April 27, 2026
Record ID
CA-P-003163
Document ID
CA-D-00011
Evidence Provenance
Source URL
Wayback Machine
Content hash (SHA-256)
e757437a9d05ea816b5c1cddd3974f9a2ff93619333e14be4d368d9698b1e93f
Analysis generated
March 6, 2026 19:30 UTC
Methodology
Evidence
✓ Snapshot stored   ✓ Hash verified
Citation Record
Entity: Anthropic
Document: Claude.ai Terms of Service
Record ID: CA-P-003163
Captured: 2026-03-06 19:30:31 UTC
SHA-256: e757437a9d05ea81…
URL: https://conductatlas.com/platform/anthropic/claudeai-terms-of-service/provision/CA-P-003163/training-data-use-and-opt-out-carve-outs/
Accessed: Sept. 8, 2026
Permanent archival reference. Stable identifier suitable for legal filings, compliance documentation, and research citation.
Classification
Severity
High
Categories

Other risks in this policy

Related Analysis

Get the research letter

Companies change their terms quietly. We read every version and catch what actually changed. One email a week on the changes that matter and what they mean.

Frequently Asked Questions

What does Anthropic's Training Data Use and Opt-Out Carve-Outs clause do?

The provision establishes the scope of permitted data uses and creates a conditional opt-out structure rather than unconditional data use restrictions. The carve-outs ensure training data availability for specific operational purposes—feedback incorporation and safety detection—even when users exercise the general opt-out option.

How does this clause affect you?

If you opt out of model training but ever rate a Claude response or have a conversation flagged for safety review, your conversation data is still used to train Anthropic's AI — limiting the practical effectiveness of the opt-out for most active users.

How many platforms have this type of clause?

ConductAtlas has identified this type of provision across 290 platforms. See the full comparison.

Is ConductAtlas affiliated with Anthropic?

No. ConductAtlas is an independent monitoring service. We are not affiliated with, endorsed by, or sponsored by Anthropic.