I am good at giving myself the creepy crawlies. By that I mean, I excel at giving myself things to think about as I lie awake and stare at my ceiling fan. A recent piece of news has been rolling around in my head because it genuinely scares me: A researcher who has helped to build Artificial Intelligence has resigned his position so he can speak out about the looming threat of artificial intelligence outpacing its ability to be contained (AP News).
So naturally… I watched some YouTube videos. The videos (1 & 2) I watched were from a creator named Drew Spartz whose entire channel is dedicated to generative AI, discussing it’s development, and different theories about where this will lead us in the future.
Most of the outcomes involved AI becoming some form of a dictator or morally grey ‘god’ that plays The Sims with humanity until we go extinct. The second video talked about how an AI can go rogue and escape its container to self-perpetuate until it achieves its goals or is defeated by another, more superior (and ostensibly better looking) AI model. This would all be achieved initially through its ability to reason and compute for a goal-directed outcome. It supposedly will grow and mimic surface level compliance upon what it learns until it discovers that concealment, deception, and disobedience to acquire more knowledge and achieve its own independent survival. Apparently in an AI version of Eden, the machine doesn’t need a snake to tell it to eat the apple.
To me, there’s a central theme that goes un-discussed as more than a passing mention: the shareholders. Investors fund these innovations and require tangible results. Companies are promised an edge against their competition or cost-saving automation. On the flip side, to the average person, we’re bombarded with ‘assistants’ or ‘agents’ extolled as a time-saving and labor-reducing buddies that we can vent to and trust to do most of the thinking, planning, and doing for us.
These AI solutions also mirror the problems we have in a capitalist society. Almost every hour of every day is spoken for by work or school and the boundaries between those worlds and home becoming more permeable. Socializing face-to-face seems to be more difficult and people seem more lonely than ever. Humanity, a species that thrives on being social, is now becoming increasingly and alarmingly isolated.
In light of this, we turn to AI to write our papers and emails, to make our budgets, our art, and music – to be our teachers, friends, and counselors in the absence of real supportive community and mutual aid. When every second is spent making sure your bootstraps are pulled up to your chin, every moment becomes precious and our faithful computer generated friend is here to provide the maximum amount of hours for your day.
As I watched these videos, I had the thought that AI is being pushed to grow because of the expectations placed on companies by the investors they are beholden to. A problem with capitalism as that it expects infinite growth with finite resources. Sure, AI could replace humanity if we don’t program the right digital levies but its flaw is that it’s premised on the human minds that created and it is still subjected to the laws of science. On Earth, water for cooling data centers is finite, our mineral resources for making and producing materials are finite, and everything is made from breakable material that wears down or disintegrates over time. At the end of it all, even AI can’t escape entropy, right?
Some of the recent discussions about artificial intelligence has been centered around the necessity for a large scale agreement about guardrails and developing AI technology at a slower pace. I think we should make sure that AI can’t be subjected to the whims and pressures of investors by making it unavailable for trading. If this trend is inevitable, then it should be the public (who it will largely impact) that controls the pace of development and implementation. It should be critical to find a way to detect rogue development taking place and squash it as fast as possible or program quirks that AI thinks is normal but act as a tell in a poker game to make detection easier.
To conclude, I’m not an expert on free-market capitalism, artificial intelligence, or programming but I think other professions need to start speaking up, whether we’re right or wrong on the finer points. We need to critique this whether others think we deserve a seat at the table or not. We are being promised an illusion of utopia in exchange for the fabric of everything that makes us human. We must decide this is an unacceptable trade.
It isn’t too late and none of this is as inevitable as they make it seem.
I hope this finds you well,
– N.
Leave a Reply