Woah, woah, WOAH. An official @FTC blog post by a staff attorney noting that "The FTC Act prohibits unfair or deceptive practices. That would include the sale or use of – for example – racially biased algorithms." ftc.gov/news-events/bl…
AND: "But keep in mind that if you don’t hold yourself accountable, the FTC may do it for you. For example, if your algorithm results in credit discrimination against a protected class, you could find yourself facing a complaint alleging violations of the FTC Act and ECOA."
This is such strong language. I literally printed a PDF in case it gets taken down. Meanwhile: kudos to Elisa Jillson and the FTC for the shot across the bow (and thanks to @marywhisner for spotting).
AND: "Don’t exaggerate what your algorithm can do or whether it can deliver fair or unbiased results. Under the FTC Act, your statements to business customers and consumers alike must be truthful, non-deceptive, and backed up by evidence." cc: @hartzog
new FTC, who dis😎
Journos: Opinions may differ, but I see this blog post as a signaling a shift in the way the FTC thinks about enforcing the FTC Act in the context of emerging technology. The concreteness of the examples coupled with repeated references to statutory authority is uncommon.
Even this language "under the FTC Act, a practice is unfair if it causes more harm than good" could signal a willingness to jettison the self-imposed strictures of the 1980 unfairness statement---which was and is contrary to Congressional intent in my opinion.
I had forgotten that 1990s (i.e., Gingrich era) Congress codified the unfairness statement. Thanks @kurtdwalters for the reminder! So, addendum: I think this codification stands against the *original sense of unfairness* and that Congress should throw it out.
Why call something "unfair" and then require a cost-benefit analysis? It's silly. Unfair means not fair, unjust, immoral. The FTC Act Congress had public policy in mind. Let's get back to that.
Reports are beginning to circulate about requirements to collect personal information for use in manual contact tracing. As many people have noted, manual contact tracing will require privacy trade offs. The question is how big a trade off and what do we get in exchange. 1/
What we get is a lot. As @EmilyGurley3---who is leading an online training at Johns Hopkins for contact tracers (great idea!)---has put it, "Contact tracers are in part detective, part therapist and part social worker.” washingtonpost.com/health/2020/05… 2/
Thousands of trained contact tracers, with access to rapid testing and backed (when needed) by quarantine authority, would make a big difference in containing coronavirus according to just about every expert I've read or talked to. 3/
To better understand privacy attitudes toward contact tracing apps, we surveyed subjects (total n=200) on two dates in early April. seclab.cs.washington.edu/wp-content/upl… (PDF) I thought some of the preliminary results were telling. 1/
Even with a hypothetical guarantee of "perfect" privacy, ~72% of participants said they were at least "somewhat likely" to download a contact tracing app. That number dropped significantly as we introduced the possibility of imperfect privacy. 2/
It mattered to participants who was behind the effort (Google, UN, etc.) and whether data was to be shared with government. There was a notable dearth of trust in the government to limit what I would characterize as secondary use or mission creep. 3/
I argue caution & humility when bringing tech to bear on the pandemic. For example, I discuss the early success but ultimate breakdown of Google Flu Trends between 2009 and 2013.
I note that digital contact tracing apps, even if well-architected from a privacy and security perspective, could wind up doing more harm than good and even empower malicious actors (a point I don't hear discussed much).
I have to say I'm really proud of @uw right now. It would be hard to enumerate all the ways this giant public research university and its staff, students, and faculty are helping, but they include developing & deploying Covid-19 tests (@UWVirology) 1/
Debunking misinformation (@katestarbird@CT_Bergstrom), providing context for emergency powers (Hugh Spitzer @UWSchoolofLaw), assessing the efficacy of containment measures (Elizabeth Halloran @UWBiostat), and so much more. 2/
Not to mention the leadership displayed by our alumni @GovInslee (@uw) and @MayorJenny (@UWSchoolofLaw) in responding to the pandemic and communicating with residents here and throughout the country. 3/
Apps that purport to track people infected with COVID-19 are a terrible idea imo for several reasons. Here are five: 1. In areas of low adoption, they will give people a false sense of security and could interfere with critical social distancing measures.
2. In areas of high adoption, coronapps will cause panic---especially if the data is visualized in such a way (e.g., heat maps) that overemphasizes relative densities.
3. The privacy concerns are considerable. To say that there are no privacy concerns because the data is anonymized and encrypted is to fundamentally misunderstand the nature of privacy harms, which are often structural in nature or experienced subjectively.