Proton has introduced AI Paper Trail, a free privacy utility designed to evaluate and expose the extensive personal data gathered during user interactions with major artificial intelligence platforms.
The tool functions by ingesting exported conversation data from platforms such as OpenAI’s ChatGPT or Anthropic’s Claude. It subsequently processes the input to generate a comprehensive personal privacy assessment, detailing specific data points and inferences that can be extracted from historical prompts. According to Proton, all uploaded archive data is permanently deleted immediately following the analysis phase and is not retained on Lumo server infrastructure.
In empirical testing involving a personal ChatGPT history restricted to the free service tier, the tool parsed the 200 most recent user prompts. The utility categorized the exposure profile under a specific privacy classification while assigning an AI Exposure Score of 58 out of 100, alongside an estimated advertising monetization value and multiple identified red flags.
The underlying analysis demonstrated how seemingly innocuous individual queries accumulate over extended periods to reveal granular personal attributes. Areas exhibiting the highest exposure metrics included user location data, specific technical preferences, financial indicators, and relational contexts. While the generated insights represent automated inferences rather than verified facts, the utility highlights the cumulative digital footprint left behind by routine enterprise and consumer AI interactions.