I’ve written about AI, mainly problems. In Extreme Thoughts on Artificial Intelligence I explored the idea that AI offers no benefit, only the jolt of the feeling of accomplishment like the jolt of euphoria of opiates and solving problems created by the system and philosophy that created it.
Let’s say I’m wrong. Let’s say AI offers many benefits I’m missing. How wrong might I be?
What difference to us would we see if we stopped using AI compared to if we kept using it?
Would we miss it? Would we think, “Darn, my life suffers for lack of AI”?
If we stopped using eyeglasses or pressure cookers, I’d miss them. They represent technologies that seem to benefit people without harming others. You could argue that without the backstop of eyeglasses, we might not allow ourselves to do behaviors that cause problems that they solve, like spending so much time indoors (I’m not a medical doctor, but I understand that not looking far away enough contributes to nearsightedness).
What properties make a technology net beneficial? Some ideas that come to mind without trying to be comprehensive:
- Not addictive or create dependence
- Using it doesn’t hurt others without their consent
- Doesn’t require pollution and depletion for each use
- Doesn’t contribute to dominance hierarchy—that is, doesn’t create or depend on resources that can be controlled with no alternative
- Doesn’t depend on scarce resources or otherwise promote imperialism
- Doesn’t separate people from each other or isolate
In this partial list I created off the top of my head I don’t claim to have created any meaningful criteria, but many technologies meet it: pencils, mirrors, wheels, bicycles, books, pillows, plates, spoons. Maybe you can tell I’m just looking around my apartment.
Still, AI violates all of them.
Does it have to? Could we develop AI that meets those criteria? I suppose it’s possible, but as we’re developing it, it’s augmenting the values of the culture doing the implementing, which values engagement (aka dependence and addiction) in its use and domination in its business model.
In the documentation I received after my colonoscopy, I saw that they used fentanyl. Most use of fentanyl I learn about is from the news reporting social, political, and cultural harms created by its use outside hospitals. Opiates produced by medical facilities used by anesthesiologists for medical purposes makes sense. Opiates produced by drug dealers for use on the street doesn’t.
AI today is a lot closer to opiates produced by drug dealers for use on the street. I’m not sure if today’s developers are closer to a lab in Mexico producing pills to smuggle across the border or like the Sackler family that lied about their product, accelerated with the support of companies like McKinsey. I’m not sure which type of drug pusher hurt humanity more overall.
Either way, it doesn’t seem clear to me that AI improves life. If it cures cancer or makes it so no one has to work so all can live like aristocrats, we’ll all still experience suffering. We’ll all still die.
What could I be missing in concluding that if we stopped developing and using AI, nobody would feel we missed anything? I’d love to learn what I’m missing, if anything. Please point me to resources showing value from AI I’m not accounting for. So far, I just don’t see what it could help or ways to undo or make unnecessary its harms.
Meanwhile, many of its benefits seem like AI itself will make obsolete, and ever faster.
