The research departments of Apple and Google have banded together to develop a collaborative touch tracking platform that can help people decide whether they were introduced to someone with COVID-19.
Touch tracking is a valuable method that lets public health officials monitor disease transmission and warn people potentially exposed to the illness so they can be screened. This does so by detecting and “following up” individuals who have come into touch with a person affected by COVID-19.
The first phase of the project is an API that can be incorporated with its applications for public health organizations. The next step is a machine-level touch tracing program that operates on an opt-in basis of the smartphone through iOS and Android.
To relay an anonymous Identity over short ranges, the program uses onboard radios on your computer — using Bluetooth beaconing. Servers are relaying the last 14 days of alternating IDs to other apps that are searching for a contest. A match is determined depending on the time spent criterion, and the distance between two computers is preserved.
When a correlation is made with another person who told the program they’ve tested positive, you’re alerted and should take action for checking and self-quarantine.
Touch tracing is a well-known and disputed technique but one implemented by health agencies and institutions working on various initiatives such as this. MIT’s attempts to use Bluetooth to create a privacy-conscious touch tracking device inspired by Apple’s Find My program are one such example. The businesses say some organizations found technological obstacles they couldn’t resolve and called for help.
A week ago, our own Jon Evans discussed the need for a larger monitoring system, along with the suggestion that you will need Apple and Google buy-in to make that possible.
Two weeks ago engineers from both companies began the project. One of the reasons the businesses got involved is that there is low interoperability on the products of various vendors between systems. For touch tracking, you significantly restrict its efficacy each time you split a device like this between multiple devices. With touch tracing, you need a huge amount of implementation in one program to perform properly.
At the same time, you are running into technical issues such as Bluetooth power suck, privacy worries about centralized data collection and the sheer commitment needed to get enough people to install the software to be successful.
Google and Apple have teamed up to build an interoperable API to address these problems, which will enable the vast majority of users to implement it if they choose.
The first iteration, a private proximity touch detection API, will be available for use in iOS and Android devices by both Apple and Google in mid-May. Apple and Google said in a meeting today that the API is a simple one and will be fairly easy to incorporate with current or scheduled software. The API will require applications to invite users to opt-in to touch tracing (the entire program is just opt-in), requiring their computer to transmit the secret, spinning identifier to devices that the individual “meets.” It will allow for tracing to warn anyone who may come into touch with COVID-19 to take further action.
The importance of touch tracing will continue past the original pandemic duration and beyond the periods when the constraints on self-isolation and quarantine are relaxed.
The second step of the project is to give the tracing device ever more flexibility and acceptance by taking it to the operating system level. No need to download an app, users can simply opt-in to monitoring right on their phones. Public health apps will continue to be supported but these would tackle a significantly wider usage range.
This process, which is planned to take place in the coming months, will allow the contract tracing device the opportunity to operate better, increasing battery life, productivity, and privacy. If it is done by the program, so any changes in these fields — like cryptographic advancements — would directly support the method.