𧠠âAI Alignmentâ isnât a problem â itâs a myth.
Come with me on a journey through ordered and unordered systems
November involved lots of carbs, gratitude, and Untangled:
đI launched the first ever âAI Reading Listâ
đïžÂ I Untangled the News, synthesizing the best academic papers and news articles I read over the month.
Want more of this piped into your inbox? You know what to do.
Now, on to the show!
Iâve written about the false promise of emergence, and why AI doomers are wrong to fear that we might build a rogue AI that is so powerful we canât control it. But there is another, much more thoughtful group of people working on âthe alignment problem,â who believe that if we try hard enough, we can align AI with our needs and wants. They too are making a mistake in their assumptions â albeit in a less glaring way. See, I donât think AI alignment is a problem to solve â I think itâs a myth. Letâs dig in.
âThe alignment problemâ was popularized by Brian Christian in a great book by the same name. Christian explains that human values, wants, and needs cannot easily align with the outputs of AI. For example, Christian asks that we critically consider ânot only where we get our training data but where we get the labels that will function in the system as a stand-in for ground truth.â Right, as Iâve written before, data reflects social biases, and the labels we use to classify people and things encode the values and beliefs of the labeler. Christian also takes the reader inside the decisions and assumptions developers make â for example, that the situations the model encounters in the real world will resemble, on average, what it encountered in training. Or that âthe model itself will not change the reality itâs modeling,â but as Christian rightly notes, âIn almost all cases, this is false.â
Furthermore, Christian warns that embracing AI inculcates problematic, predictive thinking, predicated on the idea that we can model the world. Christian writes,
âWe are in danger of losing control of the world not to AI or to machines as such but to models. To formal, often numerical specifications for what exists and for what we want.â
I agree! It would be a big problem if we started to assume that the map actually is the territory or if we let technical systems, as Christian puts it, âenforce the limits of their own understanding.â Substituting a model of the world for the real world and all of its complexity gets closer to my concern. But Christian is ultimately telling a story of progress â of how AI researchers are making small steps forward in the march toward âalignment,â as if it might just be around the corner. I donât think that it is.
Keep reading with a 7-day free trial
Subscribe to Untangled with Charley Johnson to keep reading this post and get 7 days of free access to the full post archives.