Google and Apple partnership for Contact Tracing


#61

Another important distinction is between an “app” and a “protocol”. What Google and Apple are delivering is a protocol, and presumably a central server to receive the reports. They may accompany this by an app, but it will be one among many.

Singapore seems to use something very similar to what Google and Apple are doing, but they’re also storing your phone number when you create an account. And there’s a way for the authorities to deanonymise the people you’ve been in contact with. So it’s not “privacy respecting”, regardless of what they claim (whitepaper here).

Germany has reconsidered and is now also backing the google/Apple scheme.

France is using a scheme that apparently is also not privacy protecting, but I can’t really discern that from the whitepaper. However, the people behind ROBERT have acknowledged this flaw in their response on a Github issue on the subject. I’m not sure whether France will fall in line with the rest of the EU to guarantee interoperability, or whether they’ll dig in their heels and stubbornly continue with their own incompatible implementation. Knowing the French, it could go both ways.

Australia apparently is also doing its own thing, but I haven’t found any details on their implementation in the short time I’m willing to allocate today. It appears to be similar to Singapore’s approach though, and also requires a phone number to create an account, which raises some red flags.

All of these implementations are (or will be) open sourced so the public can inspect them. All of them are, for now, entirely voluntary. And that’s unlikely to change for reasons I’m not going to repeat.

In conclusion: sure, some countries are doing their own thing, and left to their own devices, it appears quite a few of them are bungling it. That’s precisely why Google and Apple are pushing for this: in order to have one standard and secure way of doing this across the globe. Because COVID-19 does not stop at the border. Having one standard way of doing contact tracing is greatly beneficial because it would make the data relevant across borders. Because people actually do cross borders on a regular basis. Google and Apple represent the majority of mobile phone users. That’s why they are in a unique position to push for a unified and privacy respecting approach to contact tracing.


#62

Yes, that too.

That too will presumably be incompatible between the Google/Apple effort and the efforts of individual countries.

This could break the anonymity. Collecting any personally identifying information (name/alias, mobile number, age, approximate location) during registration seems like it is undermining the anonymity guarantees.

Exactly what information is sent to a central server and what information is sent to a peer remains to be seen.

Normally, yes. A country could perhaps argue that while there is any need for this app, that won’t be happening, or will be restricted to certain countries.


#63

related > https://protonmail.com/blog/truth-about-anonymized-data/


#64

If an app transmits a true random, meaningless, unique id that is generated on the phone and which changes fairly frequently - and that is all that it transmits then that is about as close to anonymized as you can get.

However

  • you are trusting that there are not Bluetooth listening devices around that record more than just the id (those devices could be modified clients or they could be dedicated devices - the obvious additional data point would be the physical location)
  • you are trusting that the app really does transmit only the id (and that it does not even have access to any PII)
  • you are trusting that the host operating system is doing Bluetooth address randomization and doing it well
  • you are trusting that your device cannot be profiled in order to reveal a little more information about the device itself (this would particularly be the case if WiFi is also enabled but potentially also via the cellular modem) - or a related similar attack where information via the app is correlated with information via one or both of the other two radios - Hardware Kill Switches anyone? - but I think the L5 does not allow separate control over WiFi and BT i.e. would have to have both “on” and then software disable the WiFi

#65

i’m more worried that untill the El-cinque Evergreen-batch comes out nobody would want to use wifi/bt around other people … :mask:


#66

WiFi has always had that problem i.e. if in a public place (or any place where not planning on using WiFi) then WiFi should be turned off. Otherwise you leak too much info. On hardware that has a Hardware Kill Switch that means using it. On other hardware that means disabling WiFi in software and hoping that it does something - and it probably does.

I always have BT off except when in the car so BT has never before been a conscious consideration for me. (Does contact tracing work car-to-car e.g. when stopped in traffic? That may be some false positives.)

However the whole COVID thing has highlighted that good hygiene extends beyond a mere virus to include wireless protocols too. :slight_smile:

You shouldn’t be sharing your viruses any more than you should be sharing your data. The Librem 5 is an anti-virus.


#67

speaking of sharing … this is definately worth it > https://lbry.tv/in-shadow:4
credits go to > https://www.inshadow.net/

beware !!! it get’s dark but also uplifting … it’s like they knew in 2018 what was coming in 2020 :mask: