Can Facebook have your data and yet not have it?

Google and Apple seem to be at loggerheads again. And this time it is not about product or patents, but Google’s advisory group asking Apple to alter its intelligent tracking protection based on which Apple is claiming that its browser is safer than any other. Safari, Apple, is contending is safer, as the AI can evolve and protect user interest. Google pointed out chinks in Apple’s claims. Apple even thanked Google for highlighting these mistakes and made changes, but Google says not much has changed as Apple has only addressed a few bugs. It believes that the algorithm that Apple is using to protect users can be reverse engineered to find the person. As users keep on defining preferences, a counter algorithm can develop a profile based on dislikes.

So, AI can be used to determine who the user is any which way. While this may seem trivial at present, imagine others trying to replicate the technology only to discover that user can be determined in another form and manner. If governments better this technology, they don’t need data on what the user likes, but what she doesn’t. Facebook, on the other hand, can be even richer. It can determine your profile more accurately based on not only what you like or love or comment on, but also what you pass or do not look at. A win-win, in either case.

Ultimately, data can be monetised a thousand different ways. So, is there a way out for the user? This is where blockchain and a decentralised AI can be beneficial. And, now some companies are experimenting with this technology to see if the user data can both be obtained and not obtained. Moving over to fuzzy logic, this would allow the company to have user data and not have user data to build profiles.

The easy way to understand is to look at blockchain. At present, blockchain works like a spreadsheet storing one information in one block. While the access to information is for both the creator of sheet and the user, without the consent of both changes can’t be made. If I were to change my name on the sheet, it won’t be possible unless the creator of the sheet assents to this change. Similarly, for companies to share my data, I will have to give approval allowing what can go and to whom.

Now, adding a decentralised AI to it will ensure that companies can extract data characteristics for advertisers by revealing preferences, but not the user. So, if a group prefers A to B, companies can sell that data to advertising without ever revealing who all belong in the group. Meanwhile, AI can create profiles for social media companies using data, but not reveal what the users like or dislike, by signaling that a set of users prefer A to B.

So, an advertisement can be both targeted and not targeted. Why would a Facebook or Amazon agree to this? While this would ensure a steady flow of advertisements for the companies, which means revenue, it shall also mean that users’ privacy stays where it is. Governments would be happy too, not too much though. As companies will still be able to block hate messages, and point inflections to authorities without revealing too much about other users.

Plus, every time a user asks to be revealed companies can share in the revenues from advertisements. But a decentralised AI will have its disadvantages as well. For an AI to perfect itself, it has to learn and learning would require it to remember certain data points. This means a trial and error method with user data. Which will have to violate some sort of user privacy.

Whether companies adopt it or not will depend on how serious users get about data privacy. Ultimately, till we don’t open the box, we won’t know what what happened to the cat.

Source: Financial Express