Privacy Nutrition Labels

My dissertation focused on designing better privacy displays for people.

Privacy policies (the corporate-and-lawyer-developed solution to stave off the threat of regulation in the mid-1990s) are long, text documents, often written with legal jargon and require a high reading level for comprehension, don’t work, and have never worked for people online.

Two initial studies on re-designing privacy policies inspired by formal labeling efforts (nutrition, energy, etc.) showed that nearly any re-design can improve privacy policies for consumers. The prior effort here was the work I won the SIGCHI Student Research Competition with, and later the ACM SRC Grand Finals (best research, graduate level).

But it was clear that adoption here would be a giant mess. Every website would need to decide to use this standard label format, or regulation would need to be created to require its use. Neither seemed likely.

At this moment though, the early app stores (Apple App Store, Android Store) were launching. These two centralized providers did have the ability to require every app on their marketplaces to use a standard privacy display, and both had already developed sets of permissions to put security controls in place. Why not privacy labels too?

We wrote two studies to understand how privacy and permissions were understood, what role they play in consumer decision making, and developed a prototype label for Android to run simulated tests on. We found these privacy labels did directly influence user decision making, though with substantial limits.

We published the work (2012 & 2013), I defended my dissertation, and the team moved on to other research ideas.

~ ~ time passes ~ ~

Then, in 2020, we (Lorrie) received a call that said within hours Apple would be launching a version of these labels. They announced them at WWDC on June 22, 2020. At the time, I was working at Google, though not on Android, and they quickly (shockingly!) figured out I was at the company, and found me to ask if I had done this work and if I had any ideas for improving upon the Apple labels. I suggested a few ideas, which as I expected were not taken up. And within months Google had announced they too would have Data Safety Labels (not privacy!) that also bore a strong resemblance to the label designs I had made in the summer of 2012. (I tell this story in the Test of Time talk that USEC awarded us for the first of the two app papers)

While this is one of my greatest stories of research impact – Lorrie, I, and our team didn’t (to our knowledge) ever directly pitch these labels to any of the right people that made this happen. We published good work, and it landed in the right places, eventually, somehow, likely with a fair bit of luck!

… and it’s a story for another time how helpful these labels are now that they have been in use for several years…