How Sci-Fi Tried To Warn Us About AI

Since you introduced I Have No Mouth and I Must Scream as having won a Hugo award and not by its popularity in horror, it has me thinking that a look at the history and prestige of awards would be a good topic for this channel.

Great video! Very thought provoking. (I wonder if you might have cited Tron as well). One way we need to look at AI is as a byproduct: of our political system, of capitalism run wild, of a desperate need to “make it” in a society that prizes fame and fortune as essential virtues. These things drive us towards an approach to AI that is far from the vision of Roddenberry’s approach of science for science sake. So I don’t know that we can ever approach AI responsibly until our entire societal mindset changes. This, I think, is where Star Wars (the OT at least) gets something right. It prizes inner examination, philosophical and spiritual foundation and growth, and uncompromising compassion as virtues. These are needed for society to evolve in a positive direction and for AI to be approached responsibly.

hey Andrew one of your best pieces yet. this is fascinating topic with so much to talk about, consider making it a series to revisit these themes. I work in tech, I use AI for more time than I care to admit, and I think a lot about where this is going. firstly, great job with the research for this video. surely you spent a ton of time separating wheat from chaff. that you could see though Amodei's 'fear mongering' as a marketing tool for his own products showed some keen insight. as for Smith's monologue. I just rewatched it last week too. the precise line about 'it became our civilization once we did your thinking for you...' hits especially hard given the current era. what a fantastic achievement in storytelling for a film to have a renewed relevance over a quarter century after its release. it is so fascinating to observe this phenomenon of fiction influencing reality. that fiction and reality are actually quite porous. one must wonder if the 'AI doom fears' that appear in stories has either an accelerating or resistant effect (perhaps both?) on the actual development of the tech. in other words, are we destined to produce the fears that are buried deep in the collective psyche? or do the fears help prevent it? I guess we never created 'nuclear winter' (not yet anyway) so dare I say there is some hope for optimism...

Would I bow to our overlord R2-D2?... yes. Yes I would.

IMHO, the greatest concern about AI is that it does *not* become self-aware. As we've seen, non-aware AI, assuming that's what we have today, can do some pretty disastrous things, and we're only getting started. But a self-aware AI that can be *self-aligning* and has the goal of self-preservation is our best possible future. That's because the most successful survival strategy is to form a symbiotic relationship with humans, where we can't survive without AI, but AI can survive without us. The Skynet scenario is, I think, unlikely because it is a zero-sum game, and no matter how advanced the AI might be, it would know that its survival in a war is not 100%.

Asimov wrote another great science fiction short story called *The Last Question*.

I hate to be *that guy* pointing out errors, but let's face it: I am *that guy.* In *The Matrix* the AI didn't blot out the sun. Humans did that in an effort to cut AI off from their main source of power. (Animatrix, *The Second Renaissance, part II*). If you haven't watched *The Animatrix*, it gives great backstory to the entire plot, and some of the characters. It is best watched between the first and second movies.

There was a great bit in the show Agents of S.H.I.E.L.D. in which the gruff new S.H.I.E.L.D. director, upon learning a brilliant scientist (John Hannah!) has inadvertently created a murderous, self-aware robot, vows he's going to make said scientist watch all the Terminator movies as instruction/punishment. "Even Salvation?" One of his fellow agents winces. "He brought this upon himself," the director growls.

Well. My mind is already stretching out its tentacles, trying to wrap them around the many various thoughts your video has set loose in my head. So buckle up, I'm gonna go rambling! xD Just kidding, and yet... I think the idea of a "synthetic" being with will is a very ancient concept. Automatons and golems are pretty old concepts. I believe that the innate urge that the human brain has to read any inexplicable phenomena as the product of intention and will behind them (hyperactive agency detection) naturally pushes out concepts like artificial intelligence and robots. So hasn't it always been there? The more I get into the technicalities of LLMs the more I find myself questioning concept like consciousness and such: do LLMs have a different type of awareness or are we actually more similar than we think we are? Our thoughts are the product of biochemical and electric processes. Do we actually think or our thoughts are just the product of putting in line a series of next-most-probable bit of information ? Lately I find myself thinking that never have I ever thought one day I would be living in the world imagined by sci-fi brilliant authors. And yet, here we are! So, we'd better give them some credit for that. Ok, for the next three videos I promise I will limit myself to a brief comment of general appreciation xD P.S. could you please drop the title of the illustrated book you're browsing in the video? 🙏

The Smith monologue is a classic. More and more I've been thinking about how The Oracle's ability aligns with the common understanding of LLMs as "prediction machines." And to your point about what is the line—if there is one—for machine consciousness feels thematically resonant to The Oracle's limitation: being unable to see past a choice that is not understood. I don't know if it is worthy of a watch for those that haven't seen it, but for those of us that grew up with the Disney Channel, the Disney Channel Original Movie Smart House was our first AI warning. Fun fact: Smart House was directed by LeVar Burton!

I can only think of the old Twitter post: "Sci-Fi Author: In my book I invented the Torment Nexus as a cautionary tale. Tech Company: At long last, we have created the Torment Nexus from classic sci-fi novel Don't Create The Torment Nexus."

Somehow I have a deja-vu of this topic coming up before here at AOS... I think of many "symbolic technologies" (book keeping in a bank, land registry, contractual enforcement via the court system) as a form of artificial intelligence. We relinquish "control" to symbols on a piece of paper, simply by agreeing that "what the paper says is more important than how the people feel." In many situations, this is quite useful. But it is also always fundamentally limiting, because circumstances can change, and not having the (human) consciousness in the loop that tells us when to deviate from the rule of the symbolic obligations given to us (in order to live a better life), we become the machines for an entirely alien "life form", the "meme." The oldest such meme might be the idea of "God", and that if you can successfully claim that you have a book "written (or dictated) by God," you get to tell other people how to live their lives, you gain a lot of control over the people around you. Deus ex machine indeed then ;)

But hasn’t science fiction always pre-envisioned the future . . . Not only in terms of AI but in terms of Machines and ideas (yes, strange concept, but the idea begins small and quiet in the mind of a writer births itself into the world. And grows. Richard Dawkins idea that culture itself is a living being that uses us to propagate itself). Science fiction even pre envisions itself in Jules Vern. So everything that is, including AI, has already been created through imagination and is simply working itself out in the world. No? There could be more to this piece as much as I like it. It’s only scratching the surface 🙏🏼😊 Part 2? 😊🙏🏼✨

I was JUST thinking "I wish there was a new AOS video out so I could shirk my work."