A Telling Future
I’m Dan Piraro, the creator of the Bizarro newspaper comic. Each week, I post my Sunday Bizarro comic, then a short essay, then the past week’s Monday-Saturday Bizarro comics written and drawn by my partner, Wayno whose weekly blog post can be seen here. I highly recommend it.
Here’s the ANSWER KEY to this week’s Sunday comic, above.
——————————————————
Osiyo, Jazz Pickles! According to a Bizarro reader, that’s “hello” in Cherokee. Thanks to Jazz Pickle Eddie G. for suggesting it.
Today’s Sunday cartoon, above, features a stereotypical cartoon fortune teller. Her client makes her job pretty easy, but lately, I’ve been anxious about an aspect of the future that isn’t as easy to predict as a cat’s next nap.
We humans have been trying to predict the future since our brains developed enough to realize there was such a thing. But however you approach it, most people agree that it’s a crap shoot.
As a rule, I avoid trying to predict the future. Philosophically, I’m all about embracing what is and living here and now, not in some more perfect world later or some bleak post-apocalypse.
But despite my resistance, a topic has got me (and millions of others) wondering hard about the future and I’m finding it hard not to want to foresee what’s coming. The topic is AI (artificial intelligence.)
It’s been in the news a lot lately and if you haven’t kept up, I don’t have the space here to fill you in so I’ll begin here. What started the wave of worry was an article last week by a journalist who interviewed Microsoft’s AI program and said the experience was deeply unsettling. I agree that the things the AI program said were profoundly disturbing. But I think a more important question is why it said it.
What I mean is that if AI can truly think independently and creatively, the way a human does, to say such things is terrifying. If it can ever actually do the things it says it wants to do, we’re all toast.
But if it is only mimicking human speech because it has been trained on billions of bits of human writing—books, scripts, song lyrics, tweets, all of Wikipedia, all of the Internet, etc.—and it cannot “think” on its own, only repeat things that it has read, it isn’t frightening at all. It is no more threatening than a child imitating something it heard on TV. Maybe it responds with threats of violence because humans are a violent species and we write about it all the time.
Experts agree that the current versions of these types of AI programs (of which there are only a few in development so far) cannot do the terrible things they may say they want to do, but what if future iterations of AI can? The possibilities are endless and apocalyptic. Is anything worth that? I’ve been waffling between these two perspectives on AI for the past week and here’s why.
I listened to a podcast a couple of weeks ago that set my mind at ease, leaning heavily toward the opinion that AI is only mimicking human speech based on examples it has been fed, can’t actually think independently, and current methods of programming will not ever get it there. (AI can also pass difficult exams like the SAT in this same way, simply spitting back knowledge it has been fed.) If you want to listen to it yourself, the podcast was The Ezra Klein Show [A skeptical take on the AI revolution] available on Spotify and probably tons of other platforms. He doesn’t include episode numbers on Spotify so look for it by the title in brackets.
After listening to that episode, I was feeling all fat and sassy, not worried about a thing. Who cares if AI has a potty mouth? Nyah, nyah, nyah. Then I read an article that was recommended by a source I trust and I got scared again.
It’s an essay by a very smart guy who doesn’t pretend to be able to predict the future but makes a couple of hugely logical points in my opinion.
First, there were numerous other human-like species around when we came along and we either out-competed them or killed them off. Adios, Neanderthals, Denisovans, Homo Erectus, etc. Given that real-world example, why in the world are we trying to create a species of being that can outcompete us? If these AI programs achieve sentience and independent thought, they will be thousands of times smarter and faster than any aggregation of human geniuses, and they’ll likely be impervious to human efforts to stop or unplug them.
Second is that it is exceedingly difficult to put toothpaste back into a tube. Once this is out there causing trouble and everyone panics, it will be too late; panic won’t help. The author’s point is that even though nothing terrible has happened yet, now is the time to panic and put strict regulations of some sort in place so that this doesn’t “escape” into the wild. As is the case with nuclear weapons, now that they exist, all we can do is try to control who has them. But imagine if nuclear weapons were cheap and easy to get, like wifi. There’d be no one left to read this post. The time to complain, sign petitions, and vote for regulation is now. At the very least, talk to your friends and family about it. We’ll need to decide as a culture if we want this.
Read his essay here but fair warning: it might keep you up at night.
I’m not normally a guy calling for more control, but I’m convinced that AI is worth controlling before it becomes something uncontrollable. I’ve no delusions that it will be put back into the tube—no profitable technology I know of ever has been—only that it will not be allowed to advance to the point the poindexters are working toward, and/or never be offered to the public via the Internet.
I’d also like to say that I hope I’m totally wrong and off base, that AI will prove to be no more dangerous than Y2K, that I’ll look back at this post in fifteen years and be embarrassed by my tone. I hope that real hard.
And finally, on a much brighter note, I’m not sure we will ever be able to program an inanimate object to think independently as animals do. That might be an unbreakable barrier, like the speed of light. We still have no idea how human consciousness works so there is no clear-cut path toward teaching a computer to do it. Mimicry may be the best we’ll ever do. Personally, I can live in a world without sentient AI so I’m hoping for the speed-of-light option. But I’m not predicting it.
I can, however, predict a smile or two as we review Wayno’s Bizarro cartoons for the week…
Incorporating adjectives into your name is one of the seven signs that you’re delusional.
Here we have a physician’s assistant in a teacup.
The first step toward becoming an artist is taking yourself too seriously. I speak from personal experience.
The comment Wayno adds to this one in his blog post for this batch is worth checking out.
Also in Wayno’s blog post this week, he tells of the bazillions of Bizarro cartoons he’s created since he came aboard our comedy zeppelin five years ago. He even posted a picture of himself with all of the original art. Time flies. Life is weird.
We’ve reached the end of our cartoon mall. Thanks for walking real fast with your elbows up high until our fitbit lets us go home. If you dig our groovy vibes and appreciate that we do this for free, without ads or pay walls, please consider helping us keep it that way via one of the links below. We will appreciate it and teach our dogs to bark your name.
Until next week, be thankful it’s this week.
BIZARRO SHOP Fun and cheap!
COMICS KINGDOM SHOP (now with Bizarro products!)
… Bizarro TIP JAR One-time or repeating. Your choice!
…Signed, numbered, limited edition prints and original Bizarro panels