Anthropic will require users of its AI chatbot Claude to decide by September 28 whether to allow their conversations to be used for model training.

The update affects Claude Free, Pro, Max, and Claude Code, while business customers using Claude Gov, Claude for Work, Claude for Education, or API access are not impacted.

Previously, Anthropic deleted consumer chat data within 30 days unless flagged for policy reasons.

Now it will retain data for up to five years if users do not opt out.

Users will see new consent prompts, with data sharing enabled by default.

The company said data sharing will improve model safety and performance. The change comes amid scrutiny over AI data practices, with rivals like OpenAI also facing legal and regulatory challenges.

Privacy experts have raised concerns about the complexity of consent in these policies.

🔗 Source: TechCrunch

🧠 Food for thought

1️⃣ Legal pressures are reshaping how AI companies handle user data retention

Anthropic’s policy shift from 30-day deletion to five-year retention reflects broader industry challenges with data management under legal scrutiny.

OpenAI currently faces a court order requiring indefinite retention of all ChatGPT conversations due to lawsuits from The New York Times and other publishers1.

This legal pressure creates a complex situation where companies must balance user privacy promises with potential legal obligations to preserve data for litigation purposes.

The timing of Anthropic’s September 28 deadline for user decisions suggests urgency driven by evolving legal requirements rather than purely business considerations.

2️⃣ AI companies face growing tension between competitive needs and user trust

Despite consumer privacy concerns, the competitive pressure for training data is pushing AI companies toward more aggressive data collection strategies.

Recent Anthropic developments