EDIT: I wrote most of this before giving a full listen to the A.I. generated “George Carlin” special. Recently I gave it a listen, and will add my thoughts to the end.
Part 1: Pre-listening Thoughts
Somebody used A.I. to “create” a new George Carlin comedy special. Kelly Carlin, his daughter, is livid. A number of people have been asking me what my take is. I’ve only heard a few samples so far, but here are my thoughts:
Like any fan, I have had that lingering thought in the back if my mind of “What would Carlin have to say about the state of the world today?” I think we can make some intelligent predictions, and AI likewise offers a way to do that in a more calculated way. However, it seems a lot of AI algorithms interject some bias of its programmers, especially political bias. I’ve seen a number of examples exposing this.
On top of this, we can only draw on his material up until he died in 2008. If Carlin had been alive 2008-2024, he could very well have ended up having a different perspective on things; that’s a significant period of time when lots of people were changing their mind on things. So again, I’m a bit skeptical of the overall predictability accuracy.
Also, Carlin’s specials typically included some topics that weren’t really topical at all. So how any AI program could predict and write a specific zany observational comedy routine on some apolitical topic (“the little world” as Carlin called it), I don’t know. I’m sure it’s possible though if you asked for something specifically. For example, “Write me a comedy bit on doughnuts, in the style of George Carlin”, and then insert that. Who knows, you might get something that would pass for a Carlin routine. (Especially given the general gullability of people who find bogus quotes attributed to him online and accept them as authentic, but I digress.) I suppose you could replace “doughnuts” in the request with “a trending, apolitical topic”. But whether anything like this was even taken into consideration, I don’t know.
But back to the political bias problem: I see a lot of people who presume Carlin would take their tribalist side on a lot of hot issues. I’m not so sure about that. For example, a lot of left-wingers hold up Carlin for his rants on religion and abortion, ignoring his condemnation of things like voting, politically correct language, and environmentalism hysteria. Conversely, I see right-wingers sharing clips of Carlin talking about “the owners” a.k.a. “deep state” of the country, ignoring all of his conservative-bashing material.
Finally, most self-proclaimed Carlin “fans” in my experience haven’t even heard most of his already-existing catalog of work. As a Carlin fanatic myself, I’m a bit disappointed when I run into somebody who says Carlin is their favorite comedian, yet has never bought a single album, book, or video of his, nor had made any effort to see him perform live back when they still had the chance. I realize that for most people, comedy typically doesn’t generate that same sort of level of devotion and support to the artist as, say, music does. But I don’t see why the world necessarily needs artificially-created Carlin when most of its target audience haven’t even heard “Class Clown” or read “Brain Droppings“.
I’ll probably get around to sitting down and listening to the whole thing, and then have more to say on it. But I don’t see the endeavor in and of itself as “blasphemy” to the work of Carlin (we’ve already seen living people rip him off) so much as something few people in the end really want and I’m skeptical could be done well. As much as I cringe at the idea, I’ll still have to listen and judge for myself.
In the meantime, as Kelly points out, it does bring up some important questions about AI and intellectual property. Should A.I. programs be barred from scraping copyrighted material and making something new out of it? And one of the questions I have: Is it fundamentally different from a living, breathing artist drawing from a set of influences and creating something new? I don’t have all the answers. But as the inherited and rightful owner of Carlin’s intellectual property, nobody should be taken aback by Kelly speaking out against, at the very least, somebody creating a product being passed off explicitly as a “George Carlin special”. Granted, I’ve been using an artificially spliced George Carlin introduction in episodes of The Devil’s Mischief for many years now, but it’s obviously no more serious than National Lampoon‘s album of spliced Nixon recordings. I’m sure however that if I were to use A.I. to create a “new Led Zeppelin song”, and then try to market it as such, Atlantic Records’ lawyers may very well knock on my door.
We’ll see where this goes. In the meantime, Carlin already left us with a wealth of actual, authentic work. So I suggest going out and buying the real stuff.
Part 2: And now that I’ve listened…
I have now taken a listen to the actual A.I. creation, “I’m Glad I’m Dead”. And overall, it was pretty much what I expected: a limited caricature of him. It was like the A.I. program was only fed the audio recordings of his special Life Is Worth Losing (2006) and maybe also You Are All Diseased (1999) and Complaints & Grievances (2001), and then a set of politically-biased Google news searches and Gen-Z Instagram posts picked the material, offset by maybe a handful of meme-driven opposing views.
To give some praise where it’s due, I think the A.I. generated some of the style and subtlety well. For example, Carlin’s occasional rapid listing of things rhythmically, or quickly rattling off statistics. Some of the topic transitions, too. The voice tone kept going back and forth between certainly sounding like a 65 year-old Carlin to unfortunately sounding like…I don’t know, some 20 year-old man.
As for the choice of material itself, I can see how somebody with a superficial knowledge of George Carlin might think it’s very much his style, but I would argue otherwise. We start with a rant about religion and monotheism, where it’s really a rehash of the same talking points we see on a typical atheism forum. This is a topic he already addressed in 1999 at the close of You Are All Diseased. So why would he just rehash the topic in 2024, especially in a more disconnected way? My guess is the A.I. was offset by the disproportionately large number of shares of the religion rants by atheists.
This eventually flows into a rant against gun ownership, which consists of just repeating the Democrat talking points we’ve all heard. It’s true that Carlin described himself as, when push comes to shove, politically left-leaning. He also took some shots (pardon the pun) here and there at “gun enthusiasts”. But this segment here was just one-way rhetoric done with a bad Carlin voice. Likewise, the calls throughout the special here and there for things like “equity” are out of place for somebody who was a shameless misanthrope, let alone fiercely against tribalism.
As I said earlier, I wasn’t sure if the whole thing would be social-political commentary (a lot of dweebs in the YouTube comments seem to be under the delusion he was a spoken word commentator and not a comedian; the reality was he was a comedian first and foremost). However, there were some segments that weren’t. This is what we would expect in a real George Carlin special. There was for example a fart joke, which, for something A.I.-generated, I was somewhat impressed by. There was also a segment on names and spelling, which I’m sure was inspired by his bits about people’s first names and his material about the English language in general.
Somewhere around the 40-55 minute segment in though, we get into something I think is more of Carlin’s style: criticizing the cowardice and incompetence of humans. There was also even some criticism about A.I. itself and self-referential material about whether the voice you hear is even real. Ironically, some of this matches the angry replies from the “George would have hated this” protesters in the comments.
I’m sure this won’t be the last attempt to create a “new George Carlin special”. Like any technology, I imagine A.I. will get better over time. We’ve already been living with A.I. in some form or another for years, whether it’s an ATM or a video game. But just as we have cinema CGI today that makes the CGI snake in Anaconda (1997) look like the low-bit pixelated snake from the Atari 2600 Indiana Jones game (1982), we may very well end up with A.I. and “deep fakes” that look indistinguishable from the real thing.
The next decade is going to be interesting. Maybe I should get to work on learning how to copyright myself.