Is it 1984 yet?
Author
Discussion

saaby93

Original Poster:

32,038 posts

202 months

Friday 10th April 2020
quotequote all
So apple and google are going to modify their devices to see who has been in contact with who.
https://www.bbc.co.uk/news/technology-52246319

hmmm scratchchin

Agammemnon

1,628 posts

82 months

Friday 10th April 2020
quotequote all
How does this sit with GDPR?

Seventy

5,500 posts

162 months

Friday 10th April 2020
quotequote all
Lol at the first in.

Mandat

4,449 posts

262 months

Friday 10th April 2020
quotequote all
Agammemnon said:
How does this sit with GDPR?
What does the article say?

GCH

4,134 posts

226 months

Friday 10th April 2020
quotequote all

Mr. White

1,103 posts

128 months

Friday 10th April 2020
quotequote all
GCH said:
Is that real? If so I'm going through my contact list tonight typehehe

Agammemnon

1,628 posts

82 months

Friday 10th April 2020
quotequote all
Mandat said:
Agammemnon said:
How does this sit with GDPR?
What does the article say?
It doesn't.

bigpriest

2,328 posts

154 months

Friday 10th April 2020
quotequote all
How does Apple or Google know someone has COVID-19? I assume it's voluntarily recorded by the user.

pip t

1,366 posts

191 months

Friday 10th April 2020
quotequote all
I can see where the OP is coming from with the thread title, but this is actually probably the least '1984' way of implementing something like this.

The use of LE Bluetooth rather than a GPS/ location based solution means nobody is 'tracking' you as such - it just requires cryptographic key exchange, and when someone has Covid 19 they basically mark their key as 'infectious.' Yes, there is the centralised cloud based database aspect, but if all it is holding is the keys, with an infectious/non-infectious marker, that's not a big problem.

Edited to add - The use of LE Bluetooth also means less possibility of false positives that could occur in a GPS based system. For example a GPS based system would probably see two people sat in two different cars in a traffic jam as two people in close proximity. Obviously there's no actual transmission risk there, but it would still flag as contact. Bluetooth wouldn't do that.

Of course, all of this requires trusting Apple/Google/app developers that the system only records what it says it does. And trusting people to log their status if infected. But all in all a pretty good solution if it can be implemented as written in that article.

Edited by pip t on Friday 10th April 20:36


Edited by pip t on Friday 10th April 20:38

Jonesy23

4,650 posts

160 months

Friday 10th April 2020
quotequote all
The problem is that if it works as described then it doesn't really work. So it can't work as described.

It's perfectly practical to make a contact tracing system, and to make it simple and lightweight. The problems are that it can't really be anonymous (for one thing, how do you push the notifications to the potentially infected?), it's likely to end up over sensitive (range/duration?) and there's no guarantee that the key data like infections will ever go into the system in the first place (asymptomatic/mild cases are still infectious but not notified).

So a nice toy, but not the most practical or useful.

Centurion07

10,395 posts

271 months

Friday 10th April 2020
quotequote all
You know the speed camera was invented by a dutch rally driver wanting to measure his speed through a corner to see if he was improving, right?

How's that working out now?

GCH

4,134 posts

226 months

Friday 10th April 2020
quotequote all
Mr. White said:
Is that real? If so I'm going through my contact list tonight typehehe
Yep


pip t

1,366 posts

191 months

Friday 10th April 2020
quotequote all
Jonesy23 said:
The problem is that if it works as described then it doesn't really work. So it can't work as described.

It's perfectly practical to make a contact tracing system, and to make it simple and lightweight. The problems are that it can't really be anonymous (for one thing, how do you push the notifications to the potentially infected?), it's likely to end up over sensitive (range/duration?) and there's no guarantee that the key data like infections will ever go into the system in the first place (asymptomatic/mild cases are still infectious but not notified).

So a nice toy, but not the most practical or useful.
In terms of the notification push, the key can be associated with a unique ID, much like the advertising identifier phones currently generate. It doesn't have to be tied to an existing email address, it can create an anonymous contact method (For example the alias contacts generated by the 'Sign in with Apple' system. Doesn't have to be tied to your real world contact details, just the instance of the software on your device.

The bluetooth aspect negates the range sensitivity issue - LE Bluetooth is very short range.

I take the point that you're relying on people to actually log their Covid status, and the non-symptomatic issue.

Jonesy23

4,650 posts

160 months

Friday 10th April 2020
quotequote all
Centurion07 said:
You know the speed camera was invented by a dutch rally driver wanting to measure his speed through a corner to see if he was improving, right?

How's that working out now?
I imagine for Mr Gatsonides it worked out quite well, financially at least.

glazbagun

15,178 posts

221 months

Friday 10th April 2020
quotequote all
Haven't Google always been tracking your android phone?

The only difference I can see is you will be alerted if you've shared space & time with someone declaring symptoms.

That's no further an encroachment of privacy, it's just letting you know you never had any to begin with.

Mandat

4,449 posts

262 months

Friday 10th April 2020
quotequote all
Agammemnon said:
Mandat said:
Agammemnon said:
How does this sit with GDPR?
What does the article say?
It doesn't.
The article says it would be a voluntary scheme, hence there's your answer on the GDPR angle.

plasticpig

12,932 posts

249 months

Friday 10th April 2020
quotequote all
pip t said:
In terms of the notification push, the key can be associated with a unique ID, much like the advertising identifier phones currently generate. It doesn't have to be tied to an existing email address, it can create an anonymous contact method (For example the alias contacts generated by the 'Sign in with Apple' system. Doesn't have to be tied to your real world contact details, just the instance of the software on your device.

The bluetooth aspect negates the range sensitivity issue - LE Bluetooth is very short range.
I wouldn't consider 400 (or 100 meters for that matter) short range. Some Bluetooth 5 chip sets claim ranges far greater than this. Measuring distance between phones using just BLE is fraught with difficulties. Problems include atmospheric conditions and physical obstacles between the two phones.

You would really only want a warning if you had come within a 10 meter radius of someone infected. You would really struggle to get this sort of resolution using all the location techniques Google use.

craigjm

20,643 posts

224 months

Friday 10th April 2020
quotequote all
This doesn’t need Apple and google it just needs the mobile phone networks. As part of the Covid lockdown in South Africa they are using mobile data to identify who a victim has been in contact with and contacting them to say “you were in the same supermarket as x who is now ill please come and have a test” kind of thing.

craigjm

20,643 posts

224 months

Friday 10th April 2020
quotequote all
Agammemnon said:
How does this sit with GDPR?
Article 9 (2) (i) of the GDPR allows the processing of special categories of data if the processing is necessary for reasons of public interest in the area of public health, such as protecting against serious cross-border threats to health.

pip t

1,366 posts

191 months

Saturday 11th April 2020
quotequote all
Slightly cartoony explainer of it from the Google blog page if anyone's interested:

https://blog.google/documents/57/Overview_of_COVID...

Link to blog page with more technical in-depth explainers:

https://blog.google/inside-google/company-announce...