No Such Thing As A Free Meal

There are always strings attached, they say. Perhaps this is nowhere more true than in the human mind, which believes itself to be in charge of its own conscious decisions when evidence frequently suggests otherwise. We, humans, tend to at least implicitly believe that we have “free will”, which is to say the ability to make choices with them being in some sense intrinsically determined in advance. Tor Nørretranders is just one of many researchers who has argued (in his book “The User Illusion: Cutting Consciousness Down to Size”) that humans have an over-simplistic understanding of their own minds, analogous to the User Illusion or simplistic mental model that most people have of their computers.

Pattern Recognition

The problem is that the brain and the causes of our behavior are extremely complex things, and the self-aware, reflective human mind is no more the master of its own fate than the ticking hands of a clock decide what hour it is. There is, however, a class of technologies which can examine the complexities of the human nervous system and draw powerful, predictive conclusions. That class is the intersection between neuro-imaging scanners on the one hand, and neural networks and other machine learning systems on the other. An artificial neural network trained to recognize and predict patterns in brain scans will frequently be able to predict when you will do or say something, even before you yourself are aware of having decided to do any such thing. Whereas the human mind sees itself only through a glass darkly, we could soon be utterly transparent to artificial minds of sufficient power.

Reading the Future on Your Friend List

Move over tea leaves, there’s a new oracular technology in town. Scanning a person’s brain for patterns of neural activity is unlikely to tell you anything about what they’ve done in the past, or are going to do even a few long seconds into the future. While an artificial neural network can in principle predict the next second or so of your behavior even better than you can, by studying your own (biological) neural net, longer-term predictions are more effective when scrying the crystal ball of social networks. Social Network Analysis is a big deal in intelligence agencies and elsewhere these days, with an increasing degree of AI-automation. It is a sobering thought to realize that where humans barely know their own minds at all, AI can know your mind by looking directly into the neural machinery underpinning it, by correlating your neural activity with your behavior, by looking at your online social connections with others, or all of the above, simultaneously.

Morally Neutral Snooping

That is the world we live in, now. One thing that hasn’t changed in this age of technological explosion, however, is that technology is a morally neutral tool. It is neither good nor bad except in the ways that it is used or applied. AI is so often depicted as a Hollywood plague of killer robots, and Big Data as Big Brother, that we can easily forget what an incredible tool of human innovation, progress, and evolution it could become. Sometimes it is good to have someone know you better than you do yourself, and that is when that person is a trusted friend. Society and industry have a responsibility to humanity, to ensure that AI and Big Data are our very best of friends, and not our opponents in a world of accelerating change.


originally posted here: