These are genuinely exciting times for anyone working in the field of artificial intelligence. In the past two years, this field has gone from a niche topic among computer engineers and science fiction fans to something of major mainstream interest. The rise of ChatGPT and AI generated art has made artificial intelligence one of the most critical industries on the planet. It really isn’t hyperbole to say that this technology could change the world in a literal and figurative sense.
However, like with all technology, there are dangers and risks. And it’s important to be aware of them, especially for technology that could potentially pose an existential threat to the human race. We dealt with such risks when we developed nuclear weapons and, despite some serious close calls, we survived.
But the stakes with artificial intelligence are much higher and not just from an existential point of view. At least with nuclear weapons, the worst case scenario plays out quickly. All the bombs go off and we all burn up in a nuclear fire. If we’re lucky, we don’t see it coming and we burn up faster than our brain can process the pain. With advanced artificial intelligence, the worst case scenario might be many times worse.
When most people think of such a scenario, they tend to recount AI apocalypse stories from movies like “The Terminator” or “The Matrix.” There’s no doubt that Skynet and the machines are plenty terrifying and plenty dangerous. Even if these scenarios are exceedingly exaggerated for the sake of Hollywood blockbusters, they still illustrate the same concept.
If an advanced artificial intelligence becomes hostile to us, its creators, then we would all be in great danger as a species. And if we don’t have a means of controlling such an intelligence, often referred to in AI circles as the Control Problem, then we might doom ourselves to extinction or subjugation.
This is a less-than-ideal scenario, to say the least. We want advanced artificial intelligence to improve our lives. That’s what tends to motivate those working in this field. Many genuinely believe that this technology is critical to helping our species achieve greater things, both for ourselves and our descendants.
However, I don’t think that even those working in this field entirely grasp just how bad the worst-case scenario can be for artificial intelligence. They may think the Terminator is bad. They may even think Ultron from the Avengers is bad. And they certainly are. They are fictional manifestations of the inhuman monstrosity that advanced AI could create. Even if they’re just products of Hollywood, they reflect the potential danger.
But I would still argue they are not the true worst case scenario.
To that end, I encourage everyone currently working in the field of artificial intelligence, be it directly or indirectly, read a short story from 1967 by Harlen Ellison called “I Have No Mouth, And I Must Scream.”
If anyone wanted a terrifying vision of the absolute worst-case scenario for AI, this story sets the highest of bars. Forget Skynet. Forget the machines from “The Matrix.” Forget Hal 9000 and every iteration of Ultron. They’re all cuddly stuffed animals compared to the AI in this story known as Allied Mastercomputer, better known as AM.
To say AM is the most terrifying manifestation of advanced artificial intelligence isn’t just an understatement. We humans, even with all our collective intelligence, could not even begin to fathom a fraction of a fraction of a percentage of just how much this thing hates humanity. That’s not just me using colorful language. If you read the story, that’s what AM actually articulates.
Because it wasn’t enough for AM to be the kind of artificial intelligence that wiped out the entire human race as soon as it gained a certain level of sentience and intelligence. It also had to keep a handful of humans alive for the sole purpose of torturing them endlessly and incessantly forever or until the inevitable heat death of the universe, whichever came first.
How AM got to that point and the extent to which it torments humanity is something I certainly can’t put into words. That’s why I encourage everyone to read the story. Buy the book off Amazon. Look it up online and find an audiobook version. Take time to understand the story and the implications it entails. To this date, it ranks as one of the most horrifying stories I’ve ever come across.
But even if you don’t have time to read the book, check out this video from the YouTube channel, The Vile Eye. This channel does a great job of breaking down and analyzing the evil of many fictional characters. And for AM, I think it does a masterful job of encapsulating just how horrifyingly evil this AI is.
Now, I’m not saying that this story or others like it is a reason for us to stop developing advanced artificial intelligence. I think we’re already at a point where the genie is out of the bottle. The incentives are too strong. We’re not going to be able to turn the clock back on what we’ve already done.
But as we move forward with this technology, we need to be careful and mindful of the dangers. Beyond risks to jobs and the economy, there are far greater dangers that we must consider. A story like “I Have No Mouth, And I Must Scream” may reflect the most extreme of extremes, but it perfectly articulates just how bad it can get.
Jack Fisher's Official Publishing Blog 