In a recent announcement, Carnegie Mellon University say they are joining forces with Princeton University to make progress on nuclear fusion using methods in artificial intelligence. They have formed the Nuclear Fusion Project, which “is the result of a collaboration between Carnegie Mellon Robotics Institute research professor Jeff Schneider, students from both the Robotics Institute and Machine Learning Department, and the Princeton Plasma Physics Lab. With funding from the Department of Energy, the team has been able to make significant advancements using artificial intelligence (AI) to help control the challenging dynamics of nuclear fusion”.
I have mixed feelings about this project. Of course, nuclear fusion on the surface sounds like a great engineering solution to the climate problem. Who wouldn’t want clean energy with less (but not zero) radioactive waste? However, deep down, I find this project unsettling, for several reasons.
First, breakthroughs in nuclear fusion would lead to larger amounts of potentially cheap energy, assuming it can actually be made cheap, which is not unlikely given that making large technologies cheaper is what global capitalism does best. This in turn makes it likely that large proportions of humanity would use this energy for nefarious purposes such as the development of dangerous artificial intelligence models. In fact, Microsoft has already said it wanted to use nuclear reactors to power its AI model training and development.
But more importantly, the development of nuclear fusion using AI further cements AI as a “necessity” of the technological machine. What goes completely beyond the possible dangerous uses of greater and greater amounts of energy is the fact that any technologies required to produce something ostensibly good as nuclear fusion will also become entrenched. That is how a lot of technology becomes mandatory and difficult to remove from society. Therefore, I object to the use of AI in the development of anything that is beneficial because AI itself is quite dangerous and facilitating its entrenchment is a horrible thing.
In fact, this is a great illustration of how we become entombed by technology: it starts out as something shiny and amusing like a chess computer. But then it is used to solve serious problems like diseases, societal, and engineering problems, and then it becomes entrenched and stuck on us.
Finally, as a generally remark on this project, I think it’s an all-around bad strategy to focus on producing cheaper energy when we should be focused on using less energy. Projects like this do two things: they not only provide new sources of energy for existing human needs, but they create new “needs” in the form of very advanced technology. And that advancement of the technological machine is unacceptable, because it is fundamentally unsustainable.
Projects like this Nuclear Fusion Project seem good because they are another in a long line of those that take aim at a single variable: CO2, and of course, we all know CO2 in the atmosphere is a horrible thing. But to attempt to lessen our impact by further advancing technology, a strategy that ultimately needs more mining, more polluting manufacturing, and more ecosystem destruction, is a bad strategy. Instead, we need to slow down technological development, reduce energy usage for wasteful and damaging things like AI, reduce material usage by reducing consumerist practices, and reduce our population.
Of course, modern technological society and mainstream environmentalists don’t want you to know any of that, and their wildest dream is to convince people that we can keep civilization the way it is by going after ludicrous new technological projects. That’s not to say that technology will play no part in correcting our problems. It might, but its role will be small compared to reduction, and without combining it with a highly respectful set of actions towards nature, it is more destructive than useful.
In other words, some technology may be necessary to transition away from technological advancement and polluting practices, but it can only work if the goal is going away from it, and not staying within civilization.
Concluding, in our current world society, if the ultimate aim is to live harmoniously with nature, then nuclear fusion AI projects won’t do a thing, except provide a new layer to the delusional thinking that technology can solve our problems. Unfortunately, such thinking is prevalent amongst many, who cling desperately to a destructive, technological machine that is at fundamentally at odds with the biosphere.