...
Brian shared COVID19 Hyperledger Healthcare SIG
Main points in talk:
- Cell phone data by itself cannot help in contact tracing and isolation, added facial recognition from surveillance cameras and credit card data among others
- A good public health system also needed with physical tracers
- TraceTogether uses Bluetooth for proximity testing- first attempt has built in privacy
- Added privacy with German app, with neutral central server
- Blockchain as a decentralized layer neutral central server.
| Rationale Conversation about blockchain |
|---|---|
| R0 - seems to be about 1.0-1.39 Goal, drive to < 1.0 |
| |
| Use time wisely If not, economic ruin Susceptible applies no vaccination in next weeks/months |
| Cellphone records are not granular enough. Try to correlate with cameras, credit card use, other private data. Make data sort of public, track rule-breakers. Force rule breakers to isolate. If found infected, repeat the process. Worked pretty well, causes problems, sustained over time = privacy breach |
| Success; however, new app due to pushback, draconian tracing is not good. They realize limiting spread to Singapore is not enough. Dependent on outside contacts - trading nation. Can't afford to close borders. Uses Bluetooth, more accurate. Limitation = more than a few think. However, limitation = strength. Limited distance, along with 'x' metrics for 'x' period of time, logs the others' ID (temporary ID). Initial lockdown of app, link phone number, and randomized ID, app or govt server creates a temporary ID (similar to a peer DID). Proximate users download from the edge, decrypted, and contracted by tracers for quarantine and test. A combination of a decentralized model (download app, turn on Bluetooth, voluntary participation) Also needs a % of the population to participate, in order for it to be effective. Three conditions above are needed. Three months ago, perhaps a few downloads. In Bluetrace protocol, being open-source, 620k downloaded so far. Privacy by design, to an extent. Link to the form with the randomized ID stored on a govt server. The app, even though it reduces correlation, still centralized management by a sovereign nation. Jim: privacy issues, degrees of success (appears at this point) Singapore homogeneous, small, can be controlled. Will it work everywhere? Will any approach be statistically significant. |
| Based on the Singapore proposal. App downloaded by choice. Temporary ID 17, Bob is ID 9 Initial ID generated by the app. The only thing stored is a push token. Temporary ID uploaded periodically to a central neutral server. No phone number stored in a central location. Temporary ID uploaded periodically. Local data encrypted. Presumably, data protection is there. Link in the German article can be translated in Google. What happens if one of those people tests positive? |
| If Alice is positive. She chooses to upload to the server. Bob is notified, he has the option to get a test at a public health center. If Bob is found to be positive, the process starts again. The app only contacts people who are in danger of infection - not a blanket order for everyone to stay home. For people who are detected to be in close proximity and in danger of infection. German proposal = GDPR = choice. Alice can choose, or not. Bob can choose, or not. Choice is paramount. People are softened up enough, a statistically significant portion of the susceptible population will probably install this app. |
| Why should we make this app the way it is and why should we publicize? Choice Various reasons in discussion to not install. More surveillance after the crisis and can be easily hacked. Data between 3/25-27 gathered together and represents the German population. Need to overcome objections - the app won't pose as much threat as people think. Blockchain can function as the central server because you can also converge data from other services and bust the boundaries of a state. No link to actual people. No central authority. |
| Concerns, surveillance Important to communicate how app protects and persuade more people to download the app. In a free society, needs to happen. |
| MiPasa references a downloadable app, data from multiple sources. |
Participant additional resources added to Chat. | good article on background for R model etc --- https://medium.com/@tomaspueyo/coronavirus-the-hammer-and-the-dance-be9337092b56 The Science article may be of interest in that it calls for an app-based intervention. https://science.sciencemag.org/content/early/2020/03/30/science.abb6936 |
...