Who needs talent when you've got AI?

First AI came for the writers, because their work was fairly easy to imitate. Then it came for the artists, quickly supplanting them (despite not being able to correctly draw hands). And now AI is coming for the filmmakers.

OpenAI recently teased the release of Sora, its new movie-making chatbot. Pika Labs just released a new version of its text-to-video tool that offers lip synching, allowing would-be Tarantinos to add dialogue to their AI-generated talking heads. There are a bunch of others.

ICYMI: And the Oscar for Best Robot Picture goes to...

With the Academy Awards ceremony fast approaching, it seems relevant to ponder whether, in a few years, there will be anyone left to fill the seats at the Shrine Auditorium. Or, at least, any humans.

It's not an unreasonable question.

Take Sora, for example. A couple of weeks ago, OpenAI's Sam Altman debuted some of the videos that have been created with this new tool, which as of this writing is not yet available to the general public. Based on nothing other than a fairly simply text prompt, Sora will produce video that looks pretty darned real, give or take a few glitches.

Here, for example, is a 17-second video created using the prompt: “A movie trailer featuring the adventures of the 30 year old space man wearing a red wool knitted motorcycle helmet, blue sky, salt desert, cinematic style, shot on 35mm film, vivid colors.”

I have questions. Is he on an alien planet or enjoying a day at the beach? Did he forget to pack a razor? Does he really expect that tin-plated waffle iron to get off the ground? Why is he wearing a tea cozy on his head?

Still, instantly going from typing a few words to producing photo-realistic motion pictures would have been mind blowing just a couple of years ago. Now we're all jaded.

Pika Labs' lip sync feature is only available to premium users willing to pony up $350 to $700 a year. [1] But for $10 a month, you can generate some funky and occasionally terrifying 15-second video vignettes. Like my truly disturbing attempt at creating a video of Senator Mitch McConnell being devoured by eagles. (Be sure to shield the eyes of any young children who may be watching, lest this haunt their dreams.)

But just because you can create movies using AI text prompts, doesn't mean you should — or that you'll be any good at it. (As my attempts at video creation amply demonstrate.) It still requires talent. So what this ultimately means is an even greater flood of mediocre-to-awful video content, coming to a small screen near you. Oh joy.

Caveman II: Unelectric Bugaloo

John Biggs, longtime technology journalist and co-founder of The Media Copilot blog, wrote an essay this week titled "Will AI Kill Creativity?"

For individual creators, the answer is probably no. AI-powered image, video, and music creation tools enable them to achieve vastly more than was possible for a single creator, at a sliver of the cost. But for teams of collaborators who bring different skillsets to a creative endeavor — like carpenters, electricians, musicians, lighting specialists, hair and makeup artists, etc — the answer is undoubtedly yes.

Per Biggs:

AI supercharges creativity, just as a typewriter supercharged Stephen King and a video camera supercharged Spike Jonze. It supercharges creativity like the electric guitar did for Jimmy Page, the way computer animation supercharged Brad Bird, and the way the printing press supercharged Europe.

But Stephen King buying a Selectric or Page buying a Fender didn’t put three hundred people out of work.

What happens to these professions when AI takes over? Who will remember how to do these things when the current generation of practitioners dies off? What will happen when everything goes kablooey [2], and we're back to sitting around a fire, beating sticks on rocks and spinning stories out of the dancing flames? [3]

ICYMI: Will the next Hollywood megastar be created by AI?

Like many, I have personally benefited from advancements in creator tech. As a lowly word slave working alone at a keyboard, I could not do what I do without a personal computer and the Internet. I am entirely endebted to the people who invented the things that allow me to make a living without ever having to put on pants. But probably half of the 100+ publications I've written for over the centuries are no longer with us — and a lot of that has to do with the Internet as well.

Democratize this

People in the tech industry like to talk about how tools like these help to "democratize" access to stuff previously only available to people who a) have buckets of money, or b) have spent years honing their craft. And that's certainly true.

But for AI to work this magic, humans — a whole lot of them — had to create the originals that AI is copying from. And what happens when there's no more original content? Will AI be forced to copy from itself?

Art could become like a document that's been Xeroxed too many times. When you make a copy of a copy of a copy, details get distorted and errors are exagerrated. Images lose resolution. Everything gets fuzzy.

It's a cliche to say that life imitates art. But when AI art imitates AI art, will it still be worth looking at?

So many questions, so little popcorn. If you like this post, please share it with your human friends.

[1] I am not one of those people.

[2] My money is on a electromagnetic pulse (EMP) caused by a massive solar flare hitting the earth's surface, frying all of our electronics in an instant.

[3] I guess I answered my own question there.