Although the USA does not have a federal privacy law, senators are getting ready to move forward with new legislation that would potentially classify AI conversations as sensitive data:
- Elizabeth Warren, the lead Senate sponsor and Mary Gay Scanlon, lead House sponsor, are aiming to introduce an updated Health and Location Data Protection Act which would ban AI companies from selling Americans' health and location data. Examples include sensitive health information, such as disease symptoms, photo uploads of medical records or prescriptions that people feed into AI bots.
- There is a proposal to provide $1 billion over 10 years to help the Federal Trade Commission to enforce this legislation.
Why is this especially relevant now?
There has been a rapid increase not only with people putting in their sensitive health data into AI chatbots to search or identify symptoms - AI companies are now actively encouraging users to merge health records and AI usage. Recently, xAI encouraged users to upload medical records to Grok, OpenAI launched specific healthcare-focused ChatGPT products and Anthropic released 'Claude for Healthcare'. The latter connects to various medical databases, clinical trial companies and provides Claude 'secure access to their lab results and health records.'
Unlike previous technology, AI-driven conversations often feel personal, making the data collection about a human being much more vast and varied. AI could collect not just health information combined with location data, for instance, but it could paint a much deeper picture of the mental health, preferences and weak points of an individual. People interact with AI assuming that those interactions stay with the company - but there is currently no guarantee that this will remain so as financial incentives are on the rise. OpenAI allowing adverts into its chats is a recent example of a third party being visible in what users otherwise interpret as a two-party conversation.
With no federal privacy law in place, people are currently relying on whatever promises companies make, with no way of enforcing or even checking if the company holds true to that promise. Understandably, as an increasing number of AI companies start to position themselves as health assistants, lawmakers are worried that this information could eventually become a commodity sold for advertising or profiling and are actively aiming to do something about it.
Should an AI company be allowed to sell your health information if you voluntarily shared it via a chatbot conversation? Share your opinion with us!
