I recently attended HOPE 26 and one of the talks I watched was on fully homomorphic encryption (and, importantly, how today it is usable for real workloads). It was a fascinating talk, and I will post a link to the video when it gets uploaded. It kicked up an idea that's been floating around in the back of my head relating to the importance of requesting removal of personal information from data brokers, and how that is generally gatekept by commercial vendors like DeleteMe and Incogni (or tedious government-provided services in the case of California's DROP). When signing up for any of these services, you're required to give up the exact information you want removed from the data brokers, making these valuable targets for attackers. Additionally, you have to pay DeleteMe and Incogni, and you have to verify your identity with Login.gov to use DROP.
You might be able to tell where I'm going with this. Why do I have to give up my data again in order to scrub it from the brokers, and why do I (generally) need to pay for the privilege? The concept crystallizing for me would definitely address the first point and might make the second less painful. If a centralized service ran a crawler indexing the data brokers' publicly posted data points, and used FHE or some other privacy-preserving technology to compare against users' identifiers without directly having access to those identifiers' plaintext, it could be run in a very different way, including potentially as a donation-based service. If this is feasible, the only data necessary to store in plaintext would be the data brokers' URLs and any associated contact information for takedowns, and a user-provided email address; this way, the user can receive an email notification with a set of data broker URLs if there are any matches. I want to look more into how creating such an index would work in practice, because I think if it's technically feasible, a deployment along these lines could be groundbreaking.
Very much interested in hearing others' thoughts on this!
