I’ve been thinking a lot about AI since I first tried having a conversation with ChatGPT in December of 2022.
Nobody knows for sure what the future holds or what place AI will have in it all. But I have a few guesses, because I can’t help trying to figure it out as well as I can.
My first guess is about the timeline. I’m guessing that for the first half decade there will be a lot of people talking about AI, thinking about AI, worrying about it, exulting over it, showing off cool tricks, but that it will not go much further than that. The force of inertia is a powerful thing; a body at rest etc. Most of us will mostly keep doing the same things we’ve been doing, sometimes with some AI-enhanced assistance. There’s also a lot of learning, tweaking, developing that needs to happen in these first few years. Scammers will get a bit more sophisticated and will have better grammar, but we’ll also get more sophisticated at spotting AI generated scams. We’ll be cautious to rely on AI too much, and rightly so considering some of its current limitations. So, no big changes. Something akin to what the past half a year has seen.
Years 6-10 will I think start to see bigger shifts. A lot of the bugs will have been worked out, and AI will have grown considerably more powerful. People will find ways to use AI that will save lots of money and be far more effective compared to human employees, and any businesses or governments that don’t adopt these emerging uses of AI will be working at a huge disadvantage. A great many jobs will evaporate, and I don’t really know where the workers will be relocated to. Many doctors could even possibly begin to be replaced, and lawyers, and teachers, and civil servants. Think of the public apprehension such changes could inspire.
I think years 11-20 is where we could see a real social transformation, though. There will be big economic questions confronting us, such as how to tax AI “workers” and how government will support what could be a growing unemployed demographic. There will probably be moral and legal questions about the personhood of artificial intelligences and their rights. There will be big questions about the use of AI in governments.
I think in some ways we’ll effectively begin to be governed by AI, at least to a point. Political campaigns will have their strategies, their slogans, their campaign promises, their policy proposals, their news releases and political speeches suggested and then written by AI. Sitting politicians will likewise come to rely on it. This could conceivably be a very good thing, although it’s not hard to imagine scenarios where it could just as plausibly be very bad.
I think the confluence of AI and robotics technology will lead to many jobs being permanently lost. At some point, probably within our lifetimes, if progress continues, it is not hard to imagine the robotic economy becoming self-sustaining; if humans continue to hold consequential positions within the way the world is run at that point, it will not be because we are needed in those positions but chiefly because we will want to have redundant safeguards against AIs being able to decide (or to be manipulated in some way) to do something that’s not in our best interest.
I also think it’s absurdly shortsighted to say, as some are saying, that human-generated artistic and intellectual content will continue to be meaningfully different from and preferable to AI -generated works of art and entertainment and scholarship. That will certainly be true for a time. But I’d guess that within two decades (and possibly much sooner), the advantage will turn decisively toward the AI. Imagine a future Netflix where the AI generates a movie never before seen by human eyes, tailored to your history of personal preferences and the mood or desires you express in the moment. The Netflix homepage will have a list of boxes to check and sliders to move, and you can tell it exactly what sort of movie you feel like watching, and how long, and you can be as specific or as vague as you want to be, and it will come up with its best guess at what you will enjoy watching. If, as you watch, you start to feel less enthusiastic about the way the movie is going, you will hit the space bar and the direction of the plot will change. Or if there’s something in particular that you want to adjust in the movie, you might hit the space bar twice and then type an instruction into the dialogue box that’ll open and the movie will seamlessly go on with the requested changes.
Likewise, rather than searching for academic papers that have been written, you’ll be able to generate a brand new perfectly argued and sourced scholarly article on any topic, that you can then cite with an easy link.
Early on, these sorts of things will be a novelty and people will try to come up with all sorts of weird and awful movies and essays etc, just because they suddenly can. Eventually, though, it’ll just be how we entertain ourselves and how we learn or argue.
This isn’t the full extent of my guesses, but these are some of the big ones that come to mind offhand. Perhaps if I think of more later I may soon do a follow-up to this post.