Massage music, AI, and psychological liberation?

Artwork generated by the Dall-E AI with the prompt “a depiction of a human experiencing psychological freedom, a gift from artificial intelligence, one-line drawing.”

I had a bizarre experience last Valentine's Day. As a gift, my wife offered to give me a massage. (While this was a welcome surprise, we're not yet at the bizarre part.) As I was settling in, I playfully asked our Google Nest Mini — it was a Secret Santa gift, and I enjoy experimenting with it — to put on some "massage music." I wasn't perfectly sure what to expect, but assumed it was going to pick some generic, soft, relaxing instrumentals. But no: the playlist kicked off with Jack Johnson, and from there careened around a space of songs that might best be summarized as "melancholy make-out."

I later realized that I was expecting the new age genre that Google calls "spa music," and was more than pleasantly surprised that what was served up was, in the end, weirder and better. In this case and in this moment in history, this musical selection was purely an accident predicated on my arbitrary selection of search term: "massage music" instead of "spa music." But as the decades roll by, these little AI's will benefit from the collective results of millions of micro-experiments, learning what music is preferred by specific people at specific times of day in specific locations under specific circumstances with specific search histories and so on. Eventually, as things seem to be tracking, these AIs will be better at picking the right music for the right moment than most of us, as individuals, will be.

And music selection is, of course, just an arbitrary little example. Gmail can now suggest how you might respond to emails — using casual responses that feel unnervingly human, since they are, in fact, predictions based on actual human responses to similar messages. This kind of technology could, presumably, apply just as well to voice conversations, making use of vocal acoustics and message content to register emotionality. As more data is collected and processed, the predictions could become more accurate and more complex, eventually taking context and conversation history into account.

Continue to walk this thought experiment further, and assume some significant but not outlandishly optimistic advances in brain science and imaging, and you eventually arrive at an AI that is aware of your sensory inputs, aspects of your physiology and brain activity, and can reliably predict your output: your behavior, your utterances, your facial expressions, your visual imagination, and even your subvocal speech — the internal monologue that registers for many people as their own "thinking," or at least the play-by-play narration of their thinking. And note that none of this requires solving the so-called "hard problem" of consciousness, which is figuring out how physical processes — electrical activity in the brain — gives rise to our mental experiences of sensing and feeling. It only assumes that we can establish a predictive mapping from inputs to outputs.

So this was the bizarre thought that occurred to me on Valentine's day: in some perhaps not-too-distant future, the robots in the room might be far more reliable predictors of my preferences, my emotional responses, and even my thoughts, than even me, their de facto "owner."

In one sense, there's nothing particularly shocking about imagining this future. After all, if you accept that minds arise from brains, that brains are made of physical stuff, and that physical stuff obeys the laws of physics, then predicting your brain activity and corresponding thoughts, feelings, and behaviors is a statistical modeling problem, not a metaphysical mystery. Nor is predicting how I will feel, better than I can, a particularly high bar as modeling problems go: many researchers have found that we are astoundingly bad at estimating how our own choices will impact us emotionally.

Rather, this little massage thought experiment held my attention not because it provided a great leap of insight, but rather because it showed a gradual path, traversable by thousands of small steps, toward a future where we realize not only intellectually but also come to experience, from — a regular, first-person point of view — that we are predictable processes: complex but increasingly transparent mappings from stimulus to response, enormously sophisticated bags of chemicals and electrical activity.

Advances in genetics, psychology, and brain science have been pointing in this direction for some time. But intuition tells me that while reading and discussing these theories engages us cognitively, it will be the daily experience of being surrounded by technology that can predict our behavioral patterns that will allow this truth to sink in on a deep, worldview-impacting, even spiritual level.

A typical response to contemplating this kind of future is to regard it as horrifying, dystopian, dehumanizing — a vision that strips us of the magic of being aware and being alive. Pervasive nihilistic meaninglessness. That may well be so, especially if technology-related psychosocial-economic trends continue in the direction of increasing isolation, wealth disparity, and lack of meaningful connection.

And yet, what if a deep realization of our physical, natural — and therefore modelable and predictable — essence could lead us toward, rather than away, from grasping our true, and liberating, essence? After all, isn't the illusion that we are somehow separate and apart from the universe and its natural laws a source of tremendous suffering? Isn't the idea that each of us is somehow orchestrating our lives from a kind of separate, individuated, non-physical plane of consciousness what contributes to feelings of inadequacy, anxiety, and shame? What would happen if our quotidian surroundings perpetually and experientially reinforced the idea that, far from being alone, separate, and always "free to choose," we are inextricably linked, inseparably interconnected parts of a natural whole?

What if our machines help us to realize that we are atoms and electricity, that we are poetic emergent phenomena, that we feel in charge but never are in charge, and what if that turns out to be, in a deeply ironic twist, spiritually liberating?

That would be kind of cool.