You signed an AI privacy policy. What did you agree to? (email.projectliberty.io)

A recent analysis of privacy policies from leading AI companies OpenAI, Anthropic, and Perplexity reveals significant concerns about user data collection and retention practices. Many users unknowingly consent to extensive data usage terms—often without reading the lengthy legal texts—resulting in their prompts, conversations, and even metadata being used for training AI models by default. Notably, Anthropic's 2025 policy change allows for user chat data to be retained for up to five years unless users explicitly opt out, mirroring OpenAI's practices. This raises alarming implications for individual privacy, particularly as the data can be shared with third parties or disclosed under legal obligations. The findings underscore a critical need for clearer privacy regulations in the AI/ML sector. A Stanford study suggests that policies should mandate explicit opt-in consent for data usage and training, especially for minors, to counteract the default data collection practices that currently dominate the industry. As users are often unaware of these implications, there is a call to innovate privacy-centric AI solutions that prioritize user control over data. The conversation highlights a broader societal need for frameworks that can restore privacy in an increasingly data-driven world, prompting users to take proactive steps to understand and manage their digital consent.