When I was watching the documentary series Life On Our Planet, I was struck by how realistic the CGI was. The prehistoric animals wandering through nature seemed as if they were being actually filmed, with both them and background details finely portrayed. I wondered how much time and effort had gone into getting it to look like that.
Life on Our Planet takes advantage of modern CGI and photography techniques that mean film shot in natural habitats, footage of animals that are real but have been transferred to a studio and sequences conjured from scratch on a computer are nearly indistinguishable. Some of the extinct land-based animals digitally brought back to life look a little like they’re hovering across the ground as they walk, and there are a few scenes where implausible numbers of dinosaurs have gathered on the same landscape for a nice photo, but we largely move smoothly between then and now.
But that was even before the rapid advances within the last year of the current generation of advanced AI software. Now there is new software called Sora that goes even further and can take a written prompt and turn it into video.
As an example, here is one prompt.
“Several giant wooly mammoths approach treading through a snowy meadow, their long wooly fur lightly blows in the wind as they walk, snow covered trees and dramatic snow capped mountains in the distance, mid afternoon light with wispy clouds and a sun high in the distance creates a warm glow, the low camera view is stunning capturing the large furry mammal with beautiful photography, depth of field.”
And this is what Sora created.
Prompt: “Several giant wooly mammoths approach treading through a snowy meadow, their long wooly fur lightly blows in the wind as they walk, snow covered trees and dramatic snow capped mountains in the distance, mid afternoon light with wispy clouds and a sun high in the distance… pic.twitter.com/Um5CWI18nS
— OpenAI (@OpenAI) February 15, 2024
Go the Sora website to see more examples of text-to-video conversion. It looks like writers can now create realistic videos without having to master the craft.
Right now the Sora videos look a little rough but you have to assume that they are going to get much better very fast, such is the speed of developments in this area.
Actor, writer, director, and producer Tyler Perry, who had been planning a major expansion of his Atlanta studio, says he halted work on it because he was shocked by the new capabilities that he felt would make much of what he had planned to do unnecessary..
Perry, whose successes include the Madea film series, said Sora’s achievements meant he would no longer have to travel to locations or build a set: “I can sit in an office and do this with a computer, which is shocking to me.”
…Perry said the breakthroughs presented by Sora would affect a range of jobs throughout the film industry, including those of actors, editors, sound specialists and transportation crew.
He said: “I am very, very concerned that in the near future, a lot of jobs are going to be lost. I really, really feel that very strongly.”
Perry said an immediate example was the construction workers and contractors who would no longer work on his planned studio expansion because “there is no need to it”. He added that he used AI in two recently shot films in which the technology was used to age his face and help him avoid hours in the makeup chair.
Who knows how far this can be taken? Entire films may one day be made by someone simply seated at their computers at a fraction of the cost. How good they will be remains to be seen. While the images may be realistic, to what extent can they reproduce quality acting?
This is a brave new world of film making.
What’s especially scary is how this technology will be exploited by bad actors (not of the theatrical kind, of course, but in the political sphere). There are plenty of people out there who want to be lied to, and this will make the lies that much more believable.
I mean, cool, I guess? I can’t help but feel the end-game here will be a complete mistrust of all forms of electronic media, everywhere. Nothing will be believable anymore unless you, personally, are able to use a loupe to detect the imprints of a ballpoint pen on a piece of ruled notebook paper written by someone you, personally, trust.
Someone made an interesting point about SORA the other day, namely that it’s sort of learning physics. If something steps in water, the water splashes realistically, things fly parabolic trajectories, etc. Sure, its just modeling billions of frames of existing video, but there’s a lot of stuff going on.
I renember when they made the “walking with dinosaurs” documentary and the CGI team sent people with big plywood feet to make proper splashes for when the dinos stepped in water. There is a whole presumed underlying physics that has to be simulated to fool us. Walter John Williams calls this “implied spaces” -- where the closer you look the more fractal detail there is, just like reality at a certain scale, rendered procedurally. Next generation games will probably do this -- knock on a given apartment door in a megascraper and there will be procedurally generated residents and their stuff and backstory. At a certain point the detail of implied spaces becomes boring.
Marcus@3- I remember a documentary talking about studios trying to make action movies more realistic by having things/people obey the laws of motion. A simple example is objects accelerating as they fall. Too realistic becomes a problem if The Hulk jumps off the top of a building and lands on the sidewalk….
Sora still has a way to go, check out the ginger cat with three front paws.
I predict the first commercial use of this will be pornography (as always). The second use will likely be studios continuing the use of the actor who grows too old to play the part (Why Captain Kirk , Why ?, (Generations)). A Harry Potter who never grew up would be a cash cow for e.g..
As to the likely harms , people with enough money could already make realistic looking CGI/fakes. This only makes it available (?) to the common person. And I’d rather be done in by the collective stupidity of humanity than the evil of billionaires/rich corporations. People can already believe that Hillary Clinton ran a pedophile ring from a pizza shop on the basis of an article someone shared on Facebook -- why would anyone need to bother with a realistic fake ?
I’m conflicted. On the one hand, I grew up putting colored stickers of dinosaurs on the correct page of the storybook and watching the very few claymation and cartoon stories about dinosaurs. I think so long as it’s presented as, “This is how scientists believe dinosaurs looked and behaved”, then not even children will be fooled that they’re seeing reality, but their experience would be breath-takingly better than peel-and-stick dinosaur stickers and cartoons.
On the other hand, I can also see a world where nobody believes anything they see/the gullible believe any stupid thing they see. Remember the commercial with Fred Astaire shilling for some current company? We are apparently living in the time of Idiocracy.
Film-making my ass — this is a brave hew world of propaganda, lies and mass-gaslighting.
(PS: that mammoth video was pretty well done; my only quibble was that the snow they kicked up would not rise that high in the air.)
Bad actors promulgating misinformation will love this. We are not far from needing complex cryptograhic security to verify every political message.
Holms, meh. So will their opponents — those countering misinformation, that is.
Here’s the thing: it is every bit as easy for anti-bad (heh) actors to do the very same.
Say a video comes out of something bad actors want to depict; next thing you know, the very same video comes out but contradicts the original video.
And of course, consilience is a thing.
—
… was gonna add something along the lines that only saps would fall for it (like, you know, the Nigerian Prince scams), but then there’s https://www.arthur-conan-doyle.com/index.php/The_Case_for_Spirit_Photography
—
BTW, in the news: Piers Morgan and Oprah Winfrey ‘deepfaked’ for US influencer’s ads
I doubt it would work, because the weakest link in any cryptographic system is always the people operating it, and pols and their lackeys are generally quite bad at even basic IT security.
However, I’m not personally all that worried about the predicted deluge of misinformation, because people already fall for hilariously obvious fakes and disbelieve well-attested and thoroughly documented facts entirely on the basis of how well they comport with their pre-existing beliefs and biases anyway. On the one hand, misinformation doesn’t need to be particularly convincing in order to be effective if people want to believe it anyway, and on the other, reality itself isn’t convincing enough to persaude people of things they really don’t want to believe.
The fundamental mistake is to believe that people (in general) come to their positions (political and otherwise) through a rational appraisal of the available evidence, when any rational appraisal of the available evidence shows that this is largely not true. If it were true, there would be a good deal less disagreement about many important matters.
You may find this of interest
https://www.bbc.com/news/world-us-canada-68440150
I do think that if people find the above genuine , Trump may lose his hard core supporters even if he gains some black voters.