1/ How can we safely reopen our cities?
You'll see many plans that rely on invigorated contact tracing to "reclaim containment" 💯
Some argue for use of apps to track our close contacts-> instant digital contact tracing
Should we be pushing for this?
science.sciencemag.org/content/early/…
You'll see many plans that rely on invigorated contact tracing to "reclaim containment" 💯
Some argue for use of apps to track our close contacts-> instant digital contact tracing
Should we be pushing for this?
science.sciencemag.org/content/early/…
2/ Let me first distinguish this from use of aggregated anonymous location data to help understand changes in mobility and the effectiveness of social distancing interventions
I think it will be feasible, useful, and I'm OK (I think) w the privacy risks
wsj.com/articles/googl…
I think it will be feasible, useful, and I'm OK (I think) w the privacy risks
wsj.com/articles/googl…
3/ I'm also not talking about apps used by public health authorities after someone's identified as a case, to monitor for daily symptoms, or even to enforce quarantine (voluntary or not)
If you have it, I'm OK w use of police powers to keep others safe
technologyreview.com/s/615329/coron…
If you have it, I'm OK w use of police powers to keep others safe
technologyreview.com/s/615329/coron…
4/ What I'm talking about is the idea of apps "keeping a record of proximity events between individuals, immediately alert recent close contacts of diagnosed cases and prompt them to self-isolate."
China's already done it, and I've seen a half dozen pitches to do it in US
China's already done it, and I've seen a half dozen pitches to do it in US
5/ I think the interest in this is partly just because there are so many brilliant technologists in Silicon Valley and elsewhere who are looking at the unfolding pandemic helplessly, and they want to help
I applaud this, but we need to consider public health utility and privacy
I applaud this, but we need to consider public health utility and privacy
6/ First, Public Health Utility
The major problem w this will be familiar to anyone who's tried to build a network from scratch- you need very high rates of adoption for the app to be useful
Even if you get 1/3 of the population to adopt it, only 9% of contacts will be covered
The major problem w this will be familiar to anyone who's tried to build a network from scratch- you need very high rates of adoption for the app to be useful
Even if you get 1/3 of the population to adopt it, only 9% of contacts will be covered
7/ So-how do you get virtually everyone to put an always-on app on their phones that tracks their contacts, eats battery, and doesn't do anything to delight them?
"App which was not compulsory but was required to move between quarters and into public spaces and public transport"
"App which was not compulsory but was required to move between quarters and into public spaces and public transport"
8/ In Singapore, there have been 620,000 downloads of the "Trace Together" app.
It's probably better than this because there will be clustering within social circles, but here's what that looks like in - roughly 1% of random contacts would be covered.
cnbc.com/2020/03/25/cor…
It's probably better than this because there will be clustering within social circles, but here's what that looks like in - roughly 1% of random contacts would be covered.
cnbc.com/2020/03/25/cor…
9/ So again, how SPECIFICALLY would the US get widespread adoption of something like this, to provide public health utility?
It would have to be some sort of automatic update pushed out to Apple/Android operating systems by default, with an opt out.
It's very tempting.
It would have to be some sort of automatic update pushed out to Apple/Android operating systems by default, with an opt out.
It's very tempting.
10/ I'm sure there are brilliant technical solutions to privacy-protections- @djweitzner is my guru in this and he and his community have been thinking about how we can have both privacy and security
pepp-pt.org is a European collaborative
pepp-pt.org is a European collaborative
11/ But the types of questions posed by @SenMarkey to @USCTO are not frivolous.
washingtonpost.com/technology/202…
If such a tool exists to trace close contacts of most Americans, what are the chances that they will not be used for law enforcement, by intelligence agencies, or hacked?
washingtonpost.com/technology/202…
If such a tool exists to trace close contacts of most Americans, what are the chances that they will not be used for law enforcement, by intelligence agencies, or hacked?
12/ @RonWyden (who has had a long history of attention to digital privacy) and @RepAnnaEshoo have said before we move ahead, we need
1) Data minimization
2) Use limitation (incl law enforcement)
3) Data Security
4) Prohibit reid
5) data destruction
eshoo.house.gov/sites/eshoo.ho…
1) Data minimization
2) Use limitation (incl law enforcement)
3) Data Security
4) Prohibit reid
5) data destruction
eshoo.house.gov/sites/eshoo.ho…
13/ To me, this is NOT just a question of "well, this lockdown sucks, if we need to give up some privacy to win the battle, so be it."
and it's not "never give up any civil liberties, you will find it hard to regain them"
It's whether the risk to privacy is worth it in security
and it's not "never give up any civil liberties, you will find it hard to regain them"
It's whether the risk to privacy is worth it in security
14/ It's stuff like this that worry me about giving up privacy for little to no gains in security
A multimillion-dollar cellphone surveillance program that sifted through metadata from millions of individual phone numbers yielded just two unique leads
washingtonexaminer.com/news/nsa-spent…
A multimillion-dollar cellphone surveillance program that sifted through metadata from millions of individual phone numbers yielded just two unique leads
washingtonexaminer.com/news/nsa-spent…
15/ What's the alternative?
traditional public health surveillance.
Trained public health epidemiologists calling lab-confirmed cases, and ASKING THEM ABOUT THEIR CONTACTS.
It would be so "us" to starve public health of resources and then crash on a useless whizbang project
traditional public health surveillance.
Trained public health epidemiologists calling lab-confirmed cases, and ASKING THEM ABOUT THEIR CONTACTS.
It would be so "us" to starve public health of resources and then crash on a useless whizbang project
16/ It's too easy to confuse sensible voluntary measures to make case finding more efficient, with borderline useless efforts that pose serious privacy risks
Poor @ScottGottliebMD got a big dose of misplaced backlash yesterday.
Poor @ScottGottliebMD got a big dose of misplaced backlash yesterday.
