Machine learning, as a topic, is a little like the dark arts in Harry Potter. When I first had the pleasure of listening to Dr. Rob Tibshirani and his then highly accomplished mentee Dr. Daniela Witten speak about Machine Learning, I was fascinated. How could an algorithm discern numbers and letters from manual script? And how could an algorithm recognize faces? The approach turned out to be much simpler than I expected. An image is gridded, and the intensity of each cell within the grid is assigned a numeric score, and differences are calculated from a set of references: the reference with the least dissimilarity is declared to be object of interest. The whole field of machine learning boils down to different approaches to defining the grid, calculating intensities, and assigning dissimilarity. This is far less interesting than the hype around machine learning and AI. So what gives?
This question remained unanswered as my wife and I welcomed our first child into the world. Then I observed something remarkable. As we took scads and scads of photos of our little one, as parents are notorious to do, Google organized the photos, and did so correctly, recognizing the face of my child over several years of development, from birth on. The algorithms are disturbingly accurate. What’s more, facial recognition is neither the first nor last step of image processing, there are a host of signed data features as well which you yourself can easily browse. For instance, when the GPS is enabled on one’s phone, the date, time, and location of the picture has been collected in tandem; Google has freely taken that information without prompting or any means of redaction. What resulted in my case is an uncomfortably detailed narrative of my child’s early life, broken down in detail by date and location.
Looking back on all this, I believe I have found the answer to my question: the rise of machine learning reflects a massive push in commercial and industry to exploit violations of privacy to the very last drop.
One might protest: but there are laws, laws, about data privacy! The attempts of legislative bodies to protect data are pathetic; mostly it’s because those writing the laws don’t understand data and how they work. And this is not a recent finding. Fifteen years ago, US senator Ted Stevens was mocked mercilessly for attempting a metaphor to explained bandwidth. To one not understanding the threat of DDOS attacks, Senator Stevens’ remark was effective in urging us to consider the importance of regulating communications. But alas, many of those who joined in laughing at Sad Ted have now become voters who still voted in senators against net neutrality, and who are less literate in digital media and rights than their predecessors ever were. I have a full lack of confidence. One might go farther to suggest there are nefarious interests at play. Unregulated industry inevitably falls this way. But, it’s sadder and perhaps more enabling to think of legislature as being incompetent rather than corrupt.
So the onus for advocacy, protection, and asserting user rights falls on us, the End User. I think it’s fair to say that we, the End Users, only see the tip of the iceberg. We don’t know what goes on behind closed doors. So we need to be thoughtful, to recognize trends, and to entertain ideas of which we are supposed to be blissfully ignorant.
Given enough time and development, disturbingly specific narratives about our life will be available to incriminate us. One might bristle indignantly at what I’m trying to imply here, the usual slogans of self-empowerment only go so far. Saying, “it’s my life, and my choice” may be well and true. The point is however that your child has not consented to sharing their data, but they are subject to the consequences of data exploitation at an unprecedented scale. And you can’t take it back.
A child whose privacy has been violated from the very moment they’re born will be the target of every predatory tactic possible. Their happiness, sense of purpose, and sense of self will be constantly under attack. Every bit of data will be synthesized not to predict their every behavior, but to manipulate it. This is something deeper than just buying and selling. I think we are falling headlong into an era of data exploitation where our freewill and creativity is held hostage and subject to threats or disclosure. I don’t presume children are any more or less vulnerable than adults, but the incessant and cumulative effect of completely mind-penetrating advertising has never before been explored, and I prefer it never be.
In the new millennium, we’ve seen unprecedented changes in the sharing and impact of so-called information. The election of the United States was barraged by foreign influence through the conduit of information. And the End Users have readily assimilated all its maligned offerings.
I’m not saying parents shouldn’t commemorate their children’s most significant moments. If I had to remark on this problem I would say: vie to buy a wonderful second hand camera and print your photos in the old-fashioned way. Some even consider scrapbooking a pleasurable hobby, especially as it gives you a chance to interact tactilely with your fondest memories. Preserve your children’s future from the All Seeing Eye.
Disclaimer: this post is a work in progress and will be updated periodically to include references, breaking information, or other minutiae.