> On 16 Mar 2020, at 01:48, Bill Cox <[email protected]> wrote:
> In fact, to the point that likely we will not build a system like the one > you suggest to track COVID19 and report to our users and health officials > when we think someone was potentially exposed. The first part of my email was explaining how to do the tracking with no privacy. This was just to make it easy to understand the objectives and to provide a simplified overview of information flow. The security architecture outlined in the published paper would make it impossible to report users. The transmitted info doesn’t have any user ID info at all. These are just random numbers labelling risky meetings. The only time that user info might be exposed is when a user is tested positive. That user may then allow the random numbers from their phone to be broadcast. When one of those numbers matches a number stored in someone else’s phone, the receiver might try to guess which person they had met recently was the infected person. However, every time that a person thought to have been exposed is required to take a test, we also issue a test request to a random person. This allows us to see whether the targeted tests are superior to random testing of people. So, a request for a test would be just as likely to be a “false alarm” as a true targeted test. The bottom line is that if you want to have a fresh health certificate you must get tested, if one fails to arrived on time. That test could be anonymous. So, the security architecture doesn’t require any exposure of identity info whatsoever. Actually, the security architecture as published requires users to voluntarily cooperate, since the health certificate transmission is not part of the design. How to compel cooperation, while not compromising privacy is an interesting security challenge. The sealed computing group could probably come up with at least a minimal solution to this problem quickly. A solution based upon current technology would go a long way toward convincing someone in big tech to get behind this type of approach. There doesn’t seem to be much chance that a vaccine will be out in less than a year, so we could easily see a few billion people with Covid-19. Even the best prepared areas are seeing around a percent case fatality rate, so this could mean a lot of deaths and probably massive economic consequences. The ad-hoc contact tracing in S. Korea and China seems to have had a very favorable effect. So, a targeted app could be highly effective. An app could be out in less than a month, if the privacy issues were not allowed to slow progress. Strong crypto-security could be released with an update. The development expenses for this type of app would be small change for big tech. Tech definitely need some good press these days. So, there shouldn’t be much hesitation, if you can reach the right people. There are already some apps being developed, but these don’t have large teams behind them, at least not yet. So, the window of opportunity for big tech is still open, especially considering the amount of promotion, etc. that would be needed to get rapid worldwide distribution. Passive location tracking is bound to generate a much higher rate of false alarms, since we need meter accuracy in the location data. Also, phones meeting in a telephone-mast signal dead zone would not be recorded. The payoff structure I suggest would motivate people to get the app on their phones and participate in active proximity determination. dss David Stodolsky, PhD Institute for Social Informatics Tornskadestien 2, st. th., DK-2400 Copenhagen NV, Denmark [email protected] Tel./Signal: +45 3095 4070
-- Liberationtech is public & archives are searchable from any major commercial search engine. Violations of list guidelines will get you moderated: https://lists.ghserv.net/mailman/listinfo/lt. Unsubscribe, change to digest mode, or change password by emailing [email protected].
