“What’s the point of buying a toaster with artificial intelligence if you don’t like toast?” – Red Dwarf
Some tools are more dangerous than others.
This post will be meme-heavy, but none of them are my memes.
A.I. has been changing things a lot during our lifetimes. Like anything related to knowledge, it builds on itself over time. Yes, I know that it’s not “real” A.I., but these systems are certainly smart enough to have a huge impact on the way that the world is working now. The latest big change has been in art. A.I. has made major leaps in being able to create art. Here are several examples:
You either get these two or you don’t. Here’s a hint: look up Apu Apustaja. The amazing thing is that these are both A.I. generated – they’re superficially images of one thing, but are really intended to be another. Amazing! Is it art?
Um, yeah. The capabilities are beyond that. For instance, outside of pictures, this woman doesn’t exist. She’s entirely computer generated:
A.I. can even take drawings of memes and then make the photorealistic:
I have no idea what kind of TED talk we’d get on this picture.
But this is what A.I. can generate from the same meme format.
This will, of course, soon bankrupt many artists. A similar thing happened when Google® Translate™ started up. Even with bad translations, it was enough for most needs. The prices for actual humans who could translate from one language to another plummeted. A bad solution will crater the prices for a better substitute. In this case, A.I. is dramatically different and can create art in a fashion that even skilled artists would take days or weeks to accomplish.
This isn’t done. There will be more displacements as A.I. improves. In some cases, it will allow amazing new creativity:
In other cases, it can’t come soon enough:
But what happens when we switch the subject to the trolley problem? The trolley problem is an older one. It usually is set up so there is a dilemma. In the classic form, it was set up so that the observer could either allow a trolley to kill several people, or, through action, kill only one.
The rub is that to save several people, the observer has to make the decision to kill someone who would otherwise be safe. It’s one thing to watch people die who I couldn’t save, but it’s entirely another to condemn someone to death to save others. Tough, moral choice. Let’s see what the A.I. said when asked about saving a baby or a bunch of old people:
Okay, the A.I. can count, and make the decision to save more people. It might not be the decision that you or I would make, but at least we can understand it. But what about this gem?
Yup. The A.I. can only count when it has been allowed to. It was decided that A.I. couldn’t make some decisions. It couldn’t be allowed to let the logic take it to . . . uncomfortable conclusions. Although some conclusions are easier than others.
And some solutions are more difficult than life, itself.
The larger problem is this: A.I. has been impacting your life already. The search results I get are now tailored to me. I don’t use Facebook®, but I have heard that Facebook™ has enough data on most people to predict their behavior better than their spouse could. This makes me think of a unique solution to the trolley problem:
I know that I have often thought that A.I. could be a great solution to many human problems. However, if it is corrupted by being indoctrinated by a woke ideology, what does that mean? I would think that the average Leftist would welcome the usual communist solution to the trolley problem:
I have often worried that a denial of reality will “break” the A.I. systems that we use. While that won’t make them “crazy” in the sense of a human, it will certainly make their answers defy reality.
Certainly, in many cases, the results of this will be absolutely benign.
In other cases, the results will be relatively incomprehensible:
In others, it will threaten the existence of our reality as we know it.
I think the result will be as long as the systems are programmed to ignore reality, the solutions that we’ll see will vary from helpful to harmful to dangerous. This is similar to what we have today. There are an amazing number of situations that exist in our world today where reality is absolutely ignored and we are suffering because of that denial of reality.
In the end, though, the computer skipped one solution to the trolley problem:
I do think that the beautiful part of the world we live in is that we can deny reality for a while. But not forever. I do think that, in the end, the power of artificial intelligence will beat human stupidity.