If you can hire Tiliy Norwood to star in your action movie at 5% of Scarlett Johansson’s fee, why not?
I bet Tilly could up her market price with an OnlyFans page. {{ LOL }}
intercst
If you can hire Tiliy Norwood to star in your action movie at 5% of Scarlett Johansson’s fee, why not?
I bet Tilly could up her market price with an OnlyFans page. {{ LOL }}
intercst
Mixed live-action - cartoon films have been around for many decades. The first example is “The Enchanted Drawing” in 1900. It’s in the Library of Congress.
The only difference is that the AI-generated actress is more realistic than the more primitive cartoon or CGI characters.
Wendy
Hollywood actors still make healthy incomes doing the voices for animated features. But I suspect that income declines in the world of AI.
intercst
There are tens-of-thousands of voice actors who could do those voices for animated features, a fear lower prices than hiring Scarlett Johansson or Idina Menzel or Tom Hanks - but the producer want that exact voice because it conveys something to the audience, not just because they can mouth the words or read a script.
AI may take over some of the anonymous voice parts, but I suspect there will continue to be a demand for well-known actors in the animated field.
Thomas Edison would have hated that, “piano banging”.
It’s not a question of talent – it’s the fact that the economic underpinnings of the movie industry are changing.
Steve Colbert cost $100 million/year to produce and generates $60 million in ad revenue.
The video game industry now has three times the revenue of motion pictures – and most of gaming’s “characters” are AI.
intercst
That’s the thing. AI can tweak the “voice” to the director/producer’s exact specification and test it against the target demographic to see if it’s profitable.
It takes a lot of the guesswork out of motion pictures.
intercst
You don’t know if a voice is going to resonate until years in. You may test a voice for smoothness or like ability or whatever, but you’re not going to get the kind of verisimilitude from research that you do from having years of “imprint” on society.
(For anyone thinks I’m going too far, I will happily acknowledge that there will be lots and lots of AI voices in the future, unions and tradition notwithstanding. But there will also be lots and lots of work for established actors - the big ones anyway - for a long time to come.)
Here’s an example. I ran an All News Radio station in Chicago. We wanted an “identifier” unlike any other in the market, and rather than a jingle package (which we used on rock stations) we hired James Earl Jones to voice it. At the time he was also the voice of CNN, those who recall around 1990 or so. “This … is CNN”. (We were a CNN radio affiliate, so there was symmetry.) We paid him $25,000 for about an hour’s work, and the right to use it forever after, of course.
Now we could have gotten any deep voice actor to give a slogan line and the call letters, but having James Earl Jones bounce it out in those tones that made the speaker rattle was far beyond what any focus group could have told us.
I don’t know if you have to have Ellen DeGeneris or Tom Hanks or Robin Williams or Jim Carrey or Duane Johnson in an animated movie to make it successful, but they must pay those big bucks for some reason, no?
I wonder if the part of the acting world that’s going to be affected by this isn’t the “above the title” group in major motion pictures.
Instead, I can see this tech more easily replacing other actors. The ones who have brought to life such unforgettable roles as “Girl at Supermarket,” and “Waitress,” and “Student #3.” The ones who have one line, no name, and 11 seconds of screen time. Or the ones in commercials, or low-end TV shows, or who do all the other little video chunks that show up in places like gas station screens, hotel tv loading screens, infomercials and safety videos, etc.
These are jobs that go to anonymous working actors today (who are union members) but might be replaced by an AI.
Maybe, but those people are working for union scale, meaning they get a minimal amount of money anyway. I would guess that the cost of creating a(n anonymous) character would be greater than just hiring one of the many hopefuls out there who are thrilled to get 11 seconds of screen time.
There is a day rate of anywhere from $100 to $3000 (for a known actor), and the weekly rate (last time I saw it) was around $1500 minimum. Obviously that can go much higher in the case of “star power” actors, or even in roles which the director/producer thinks vital, but not necessarily box office draw material.
The real savings would come by not having to pay Tom Hanks $20 million plus points, but again you don’t get Tom Hanks for your movie, so other than a few news stories (first time novelty) you also don’t get Tom Hanks on the talk show circuit or in the paparazzi press.
Actors actually do need actual human actors to act with, even when doing weird blue screen stuff.
And you probably can’t get an AI to do what a Tom Hanks does in a movie. After all, there are countless human actors that can’t do what Tom Hanks does in a movie. It’s hard to have on-screen charisma, comic timing, emotive ability, etc. AI can (perhaps) do a mediocre emulation of a plain-vanilla human, but probably isn’t anywhere close to actually doing a good job of acting.
Which is why I think the first major use of AI actors (if we’re ready for them) is for all the filler stuff. The one-line no-name roles, commercials and training videos, minor characters in low-end TV shows, etc.
It the same way the very first robot cooks are being used in fast food places replacing minimal-skill line cooks, not Michelin-star chefs. Or that AI slop is filling up the dregs of online news sites and the remainder bin at Amazon.com, rather than replacing the Booker prize authors out there.
I wonder if they’ve done it already? How would we know if a character in a commercial is real or AI anyway? It’s not like there are credits shown anywhere.
Yeah maybe, but as I say, those are cheap. You know what’s not cheap? Big sets, backgrounds, choreography, roaring crowds, racing cars, airplanes zooming by, that sort of thing.
Think of the Jardiance commercials with 20 background dancers all cavorting around a splashing fountain in brightly colored costumes. The crowd scenes of a beer commercial along the coastline of El Salvador. Apple’s “1984” commercial with row after row of costumed conformist masses.
Those sorts of things are what AI will be doing: eliminating the expensive sets, dozens of actors, and all the so-called “scenery” of such efforts. Instead of a 3 camera film crew on holiday in a foreign land, complete with directors, make-up, cinematographers, it’ll be one or two primaries against a green screen, and a clutch of programmers in cubes filling in the backgrounds.
Yes there will be a few exceptions, but mostly if you want to eliminate the production costs, you use AI, the same as Hollywood is using CGI for expensive war and destruction scenes. You’ll still pay Nicole Kidman or whoever if you want that “endorsement” feel.
“Will be doing?” As you point out, AI’s been doing that for years. We’ve just called it CGI instead of AI.
I don’t think we’re actually close to it being economical to have AI actors instead of human actors in “ordinary” contexts (Spider-man’s been a CGI sprite since the Sam Raimi days). I’m just saying that when the day comes when that changes, the AI will probably come first for the minor characters rather than the above the title folks - just like it came first for the extras and Marvel Third Act disposable armies via CGI….
Not quite the same. CGI has involved rooms full of artists, coders, and directors (or at least assistant directors.) AI will involve one guy shouting into a microphone something like “Make the jet do a loop de loop and then fire and blow up the Empire State Building” and it will happen, at least digitally, from which iterations or alterations can be made.
In CGI, the one “Let It Go” sequence from “Frozen” took 50 effects and lighting specialists, and over 4,000 computers logging over 30 hours each to produce 3 minutes of animation. With AI it will require fewer humans and perhaps even more server time. It worked out to 150,000 hours of server time as it was!
I think that overestimates the capacity of AI, and underestimates improvements in CGI that have been happening continuously over the last several decades before we started calling all software AI. That’s been going on incrementally all the time.
Current big-budget feature-film level CGI involves rooms full of artists, coders, and directors - because that’s what’s needed to produce top-level visual effects. And that’s not going to change right away, BTW: AI isn’t going to make that level of people any time soon. But the number of coders and artists and directors for any given level of CGI has dropped precipitously. What might have taken that room full of people in 2000, like this scene in Deep Blue Sea, could be done today with very few people at all - even using the tools of a few years ago before everything was labeled AI:
AI’s just the next iteration of CGI - which itself displaced all the people whose job it was to hand draw animation and effects with people whose job it was to use computers to generate those images.