Anthropic is implementing significant changes in its data handling policies, requiring all Claude users to decide by September 28 whether they want their conversations used for AI model training. This shift marks a departure from Anthropic’s previous stance, where consumer chat data was not utilized for training purposes. Now, the company aims to leverage user conversations and coding sessions to enhance its AI systems, extending data retention to five years for those who do not opt out.

• Anthropic’s New Data Policy: What’s Changing?
• Behind the Policy: Rationale, Competition, and Industry Scrutiny
• The Challenge of User Consent and Regulatory Oversight

Anthropic’s New Data Policy: What’s Changing?

Previously, Anthropic assured users that their prompts and conversation outputs would be automatically deleted from its backend within 30 days unless retention was legally or policy-required, or if the input violated its policies, in which case data might be retained for up to two years. This new policy affects users of Anthropic’s consumer products, including Claude Free, Pro, and Max, as well as those using Claude Code. Business customers using Claude Gov, Claude for Work, Claude for Education, or API access remain unaffected, similar to how OpenAI protects enterprise customers from data training policies.

Behind the Policy: Rationale, Competition, and Industry Scrutiny

Anthropic frames these changes as a matter of user choice, suggesting that by not opting out, users will contribute to improving model safety and accuracy in detecting harmful content, while also enhancing future Claude models’ skills in coding, analysis, and reasoning. However, the underlying motive likely involves the need for vast amounts of high-quality conversational data to maintain competitive positioning against rivals like OpenAI and Google.

The changes also reflect broader industry shifts in data policies, as companies like Anthropic and OpenAI face increased scrutiny over data retention practices. OpenAI, for example, is currently contesting a court order that mandates the indefinite retention of all consumer ChatGPT conversations, including deleted chats, due to a lawsuit filed by The New York Times and other publishers. OpenAI COO Brad Lightcap criticized this order as a “sweeping and unnecessary demand” that conflicts with privacy commitments made to users.

The Challenge of User Consent and Regulatory Oversight

The rapid evolution of these policies has led to confusion among users, many of whom remain unaware of the changes. As technology evolves, privacy policies are bound to change, but the sweeping nature of these updates often goes unnoticed amid other company news. For instance, Anthropic’s recent policy changes were not prominently highlighted on its press page.

Existing users face a pop-up with “Updates to Consumer Terms and Policies” in large text and a prominent “Accept” button, with a smaller toggle for training permissions set to “On” by default. This design raises concerns that users might inadvertently agree to data sharing without realizing it.

The stakes for user awareness are high. Privacy experts have long warned that the complexity of AI makes meaningful user consent nearly unattainable. The Federal Trade Commission (FTC) has warned AI companies against changing terms of service or privacy policies surreptitiously or burying disclosures in fine print. Whether the FTC, now operating with only three of its five commissioners, is actively monitoring these practices remains an open question (FTC warning on AI data privacy). For more information on data retention policies, you can refer to industry standards and guidelines (Data Protection Commission Ireland).

Anthropic’s new data policy for Claude users represents a significant shift, requiring an active opt-out for conversation data to avoid being used for AI model training. This move highlights the growing industry need for high-quality data and the ongoing challenges in ensuring clear user consent amidst evolving privacy policies. As AI companies navigate competitive pressures and regulatory scrutiny, the onus remains on users to stay informed about their data rights and choices.

Frequently Asked Questions

What changes is Anthropic making to its data handling policies?

Anthropic is requiring all Claude users to decide by September 28 whether they want their conversations used for AI model training. This marks a shift from their previous policy where consumer chat data was not used for training purposes.

Who is affected by Anthropic’s new data retention policy?

The new policy affects users of Anthropic’s consumer products, including Claude Free, Pro, and Max, as well as Claude Code users. Business customers using Claude Gov, Claude for Work, Claude for Education, or API access are not affected.

What is the retention period for data under Anthropic’s new policy?

For users who do not opt out, Anthropic will retain data for up to five years. Previously, data was automatically deleted within 30 days unless retention was legally or policy-required.

Why is Anthropic changing its data policies?

Anthropic aims to leverage user conversations and coding sessions to enhance its AI systems, improve model safety, and maintain competitive positioning against rivals like OpenAI and Google.

What concerns have been raised about Anthropic’s policy changes?

Concerns include the possibility of users inadvertently agreeing to data sharing due to the default settings and the complexity of AI making meaningful user consent nearly unattainable.