A little over a year ago, at the Worldwide Developers Conference (WWDC) 2016, Apple announced that it would use a technique developed about a decade ago, called ?Differential privacy? (Differential privacy).
Believe you at Apple or not, the company has proudly boasted for years that it has not violated the privacy of its users in order to obtain data that will help it improve products / services. And this technique comes precisely to "solve" this, since it works in a totally anonymous and still optional way.
The use of the technique by Apple started covering only four areas and, in iOS 10.3, it started to also include (again: optionally) data from iCloud.
Now, tell the Wall Street Journal (closed matter for subscribers), Apple is again expanding the use of differential privacy to also include history of web browsing and general health data.
But one thing that I found particularly interesting, in the report of WSJ, it was this little chart that explains well how differential privacy works:
They basically use a fictitious example of a 100-resident condominium where you need to do research using differential privacy. Of the 100 questionnaires, 10 have the question "Do you smoke marijuana?" exchanged for ?Throw a coin and answer? yes ?if it heads.? That is, imagining that the questionnaire for example also asks the color of the resident's car and that there happens to be only a single blue car in the condominium, there is no way to know if the owner of the questionnaire received the changed or not.
And it's not just Apple; companies like Microsoft and Uber (and even Google, to a lesser extent) have also been experimenting with differential privacy, not least because, if you're going to ask users, I even imagine that many would say they don't mind having their privacy violated in favor of better products / services, but I'm sure they would prefer that it wasn?t necessary.
For those interested, the settings related to this, on iOS, are within Privacy Settings Review:
And you, how much do you care about that?