Questions from BNR-Nieuwsradio's "Ask me anything" program.
Opened this issue · 3 comments
You can listen to the program (in Dutch) here: https://www.bnr.nl/player/audio/10151544/10394800
We need to have solid answers to these questions:
-
What about groups of persons that share a single phone?
-
What about persons that do not have their phone on them all the time?
-
What is the exact purpose of this app?
-
What consideration has to be made between privacy and health?
-
Do we use GPS-tracking?
-
How much insight does the government have into my history?
-
How reliable are these apps?
-
Are health and privacy connected?
-
If we use bluetooth, what do we do with our infected neighbour?
-
How is this different from a having people wear a star?
-
What about the use of bluetooth and the sniffers that are already installed for citymarketing purposes?
-
Are numbers assigned by a central authority?
Short answer: No, your phone generates one. -
What additional data can be learned from the data collected on the back end?
-
What about ethical questions like: Can my employer require me to show my status?
-
Is location data stored on the backend?
-
Do we have to log IP-addresses of all infected persons?
-
Supermarket example: A couple of days ago I came into contact with an infected person in a shopping queue...
-
Wouldn't an ankle-monitor be a better solution?
17 and 18 are answered by Bert in the program.
We need a clear written answer that communicates the same for this.... -
What are the main positive reasons to use this app?
-
How can the app be abused?
-
Can jokers register themselves as infected?
-
What is the challenge for app-builders?
-
Can the app be internationally coupled?
Short answer: Yes! Every government should run their own backend server. -
Can people be fined or prosecuted based on the app data?
-
Can the data from this app be matched to my financial or fiscal data?
Short answer: Not with the data from this backend proposal.
I think we need to group relevant questions together and give a short "one-line-answer" and a more elaborate longer one in a format like this:
Short answer: One line answer
Long answer: Multiple lines long answer
P.S. From the contents of this radio program it becomes clear that a lot of people are reading what is being written here.
To those people I'd like to say: "Please give us some credit and join us! We would much rather work with you than in parallel of you, or in competition with you! We are all in this same situation which we desperately want to get out of!"
If we decide to build this system and turn it on until a vaccine exists, we need to have constructive, but also destructive input (so we can see if we can fix it) from as many people as possible!
That's a lot of questions :) People are, rightfully, worried about this, particularly the privacy aspects. A concern that we all share.
I've made a first attempt at answering some of the questions, but definitely we need to cooperate to refine this further. Comments and rephrasing is welcome!
1. What about groups of persons that share a single phone?
This won't work, everyone needs to have their own bluetooth device that can generate keys. A technical option could be to introduce simple beacons that can generate keys that can later sync to a shared family telephone. [I am not sure how feasible beacons are - can someone comment?]
2. What about persons that do not have their phone on them all the time?
This only works if people do carry their phone, or bluetooth device, when they go outside their house.
3. What is the exact purpose of this app?
[this is an important one, so this likely needs a bit of refinement]
The purpose is to give the user(s) of the app an indication of the risk they have been exposed to so they can act on this information and take action to curb the further spread of COVID-19. Furthermore, it could help speed up the return to a more regular way of life, by allowing more movement of people while simultaneously enabling quickly curbing outbreaks.
4. What consideration has to be made between privacy and health?
Privacy and health are not in conflict here as infection reports cannot be traced to a specific person. Think of it this way: it's like Hansel and Gretel: you leave a trace of breadcrumbs wherever you go, but no one knows it was you who left them there. They only know that they saw a particular colour of breadcrumb along a path a while ago and that someone has told them to self-isolate if they've seen that colour in the past.
7. How reliable are these apps?
No app is hundred percent reliable. However, no intervention or test is hundred percent reliable either.
9. If we use bluetooth, what do we do with our infected neighbour?
If you indicate your 'home' location, an app can, without transmitting your actual GPS coordinates, keep track of frequent nearby signals that can be excluded. Another more pragmatic option is to deactivate contact tracing altogether as long as you are within a several meter radius of your home [I think this unlikely to be a big challenge, as the effective range of bluetooth, through walls, is rather limited - thoughts?]
10. How is this different from a having people wear a star?
[This question by itself is a rather distasteful reference, let's say: 'a badge' instead]. A badge would only work, if you could travel back in time to apply it to yourself once you find out you are diagnosed. The main problem is the delay between being infected/infectious, getting symptoms and being formally diagnosed. A contact app can more or less do the 'travel back in time' part, informing those you've been close to in the past, so they can take proper action.
16. Do we have to log IP-addresses of all infected persons?
No, we work with an intermediate representation that can't be traced back to a specific IP address.
21. Can jokers register themselves as infected?
See it this way: just like you can't get medication from the pharmacy without a prescription, you won't be able to register yourself as formally infected without it being signed off by a medical professional.
That's a very nice first attempt!
I would like to comment on question 9; the neighbour issue.
I don't see the real issue here actually. It should be possible for the app to specifically show you when you where close to an infected person. When the app tell you: you where close to an infected person the last three days except the timeslots that you where not at home. Then you are able to reason back in time. If it was not your partner sitting next to you on the couch you can assume that it's (one of) your neighbour(s). Should you place yourself in quarantine? That only depends on how close your relation is with that neighbour ;) But if you didn't had physical contact you probably don't have to go into self quarantine.
And as long as the app can give you quite specific timeslots you can always reason back who it might have been.
Is this a privacy concern? Maybe.
9. If we use bluetooth, what do we do with our infected neighbour?
If you indicate your 'home' location, an app can, without transmitting your actual GPS coordinates, keep track of frequent nearby signals that can be excluded. Another more pragmatic option is to deactivate contact tracing altogether as long as you are within a several meter radius of your home [I think this unlikely to be a big challenge, as the effective range of bluetooth, through walls, is rather limited - thoughts?]
This is a nice idea, but it has the drawback of having to enable location services (because GPS won't work indoors), which is something people might not want to do.
But the properties of RF-transmission might resolve this problem for us, because in my preliminary tests with an iPhone 7 and an Android device, told me that most of the time, you will have an RSSI of above -75 if the two devices get within a 2 meter range. (See: https://www.bluetooth.com/blog/proximity-and-rssi/)
Any kind of barrier, except for glass and plexiglass, will significantly bump down the RSSI value, so this might not be a problem.
Short answer: The distance measurement used in Bluetooth works based on signal strength,
The signal strength drops considerably if obstacles are in the way and therefore your infected neighbor or housemate in a different room, will not be counted as positive.
You'd need to be carefull around glass windows and plexiglass barriers.