Facebook knows an awful lot about you; this much is obvious by now. Every message or file sent using one of its many platforms (including Instagram and WhatsApp), all of your contact lists’ details, all your log in locations, all the data associated with any apps you’ve ever had connected to your Facebook account, your political affiliation, your religious beliefs, your sexual orientation, and far more – it’s all in their hands.
It’s not altogether shocking then, as highlighted by an excellent piece in The New York Times (inspired by this thread) that the company has patents and patent applications for some rather unnerving technologies. It doesn’t mean that such tech is currently in use or is even close to existing, but it’s interesting to note what’s been proposed or, in some cases, legally snapped up by the social networking colossus.
These applications include an algorithm that uses your phone’s microphone to listen to your surroundings and absorb the electrical interference pattern being emitted by your TV set to try and work out what you’re watching or what adverts you were listening to. Another patent describes tracking your daily routine – where you eat, buy coffee, walk, and sleep – and provide notifications to others if deviations occur.
Yet another suggests looking at your online activity in order to ascertain personality traits. Whether you are more extroverted or introverted, conservative or open-minded, emotionally stable or not could potentially be revealed by nothing more than your texts and likes.
The one that’s arguably the most striking of the lot, though, is the patent application – which means it has not been granted yet – that essentially aims to predict your future. Using your online data, Facebook’s patent explains that it could be possible to approximate when you or your network members may graduate, get engaged, divorced, hired, fired, have a child or, well, when a death will occur.
The patent itself isn’t that complicated. Using automated machine learning, this hypothetical software will look at how other people’s lives are turning out while processing all your data. When a major life event happens, it will see how closely it’ll match up to any predictions it makes, adjust its calculations, and keep at it until its accuracy improves.
Why would Facebook want to do this? Well, a clue lies at the end of the abstract: The system “provides advertisements to the user responsive to the prediction of one or more life change events.”
Artificial intelligence (AI) has certainly come a long way in recent years, and pattern recognition is certainly its forte. From recognizing skin cancer samples and viable embryos in IVF to playing ancient board games and reading emotional changes, AIs are certainly very good, and sometimes better, than their human masters.
AI can, of course, also be used for more morally ambiguous purposes. It was, for example, recently suggested that AI could be used by governments to spy on their enemies better than their enemies can spy on them, and that this could start a new type of globally destabilizing arms race.
It’s not surprising that conglomerates like Facebook are investing in, or at least investigating similar forms of, data-crunching AIs or algorithms. Regardless of whether this bothers you or not, though, based on how much data we willingly hand over and how far AI has already come, the tech behind it is clearly not exclusively destined to be science fiction.
Facebook already knows you better than your friends do. It’s probably not lunacy to suggest that, with all its resources, Facebook's algorithms will someday understand you almost as well as you yourself do, maybe even predicting upcoming life events before you're consciously aware of their approach. A brave new world awaits.
[H/T: NYT]
1