The world sure is changing, huh? What a crazy place we live in.
Used to be that you could dream of robots doing cool stuff and now you can outsource your dreams of cool stuff to robots, thank god. What a time save.
AI, as I have been assured over and over again, is coming for creative jobs, and we are permanently on the cusp of a full-on AI revolution where you can watch a custom version of the Lord of the Rings but with all the Tom Bombadill content re-added with a simple click. These promises, these industry-agnostic promises, have been frustrating.
They’ve been frustrating because, and I think this is the most neutral I can possibly be about it, they are almost always being made from positions of spiteful, self-interested bad-faith. Get another job, start learning how to be an electrician, this will stop traditional gatekeepers, look at the objective quality of this generated art vs human art, humans don’t stand a chance, look at the objective quality of this generated woman vs regular human woman, women are finished etc. etc. etc.
It has been a semi-permanent drumbeat of nasty speculation about what is inside the black box and what the black box will eventually be able do, and I struggle to distance myself from this when thinking about AI at all. It repulses me. But, as I keep on being told, it is coming, it is here, so I thought I’d try my best to have a look at what’s being produced in the film industry and try and honestly evaluate it.
Getting this out of my system now, purging my biases, moisturising, calming myself.
What I want to do here is relinquish (most) of my priors and most of the existing pro-AI arguments in favour of future content to have a proper look at what is happening in the present. To (semi-) seriously look at the visible aspects of the AI film industry and assess it for what it is capable of showing us right now without the speculative nonsense and without my deeply held cynicism towards it. It is, as I have heard, gonna take over the industry and every other industry, so it has to be producing something good right? Right?
Part 1: The Enthusiastic Solo Creators
The gatekeepers are dead, long live the gatekeepers (buying tokens for midjourney or whatever). We’ve ushered in an era where we can finally make creative works from our bedroom or garage, or even at our personal computers. Wow!
The first person we're going to look at is PJ Aceturro, a name that I had encountered briefly with Princess Mononoke live action. Turns out he made an advert for the NBA and also produced the video What If Studio Ghibli Made LOTR?
What can I possibly say about this viral hit? I watched it, and it certainly does mimic the Fellowship trailer shots in a wildly inconsistent style. Not exactly a thrilling review, but what else am I supposed to go on here - this isn’t a preview of something, it’s a direct copy of a preview. As a piece of art, as a creative work, it doesn’t demonstrate anything at all.
PJ (we’re friends now, I can just use the start of his name) did make an original AI short film titled 4 minutes to live, a little bit tight as far as runtime, and for sure nowhere near his most viral work. But finally! A thing with a story! A plot! Stuff to dig into!
It’s bad! The narrative is directionless, more in love with the ability to showcase shiny new locations than in doing anything. We jump from same-feeling shot to same-feeling shot, the characters on the screen staring directly down the lens of the camera with dead eyes before looking away, guilty for wasting even this small amount of our time. Our main character is trapped waking up in a new person’s body every day (something that works to the massive benefit of the tech being used) and that this experiment in sleep…something has separated him from his partner. He contemplates suicide after a thousand different lives, but then feels the responsibility for these other people weighing on him before chancing upon a cafe in Paris and meeting his soulmate again, love winning in the end despite the people fundamentally being different.
I added the thought bubble for fun, but it does just cut from the pills cabinet to this shot lmao. Does anyone know which city he might be in???
Our main character swears to Quinn (his soulmate, not some sort of new deity I’ve made up) to always meet her here no matter what, before being flung back to the framing we were introduced to at the start, waking up inside a criminal condemned to death in four minutes (lmao) unable to get to sleep.
Like an attempt was made, for sure. But the longer we stayed with the same character, the more their face warped and changed - a metaphor, perhaps, for the foggy fallibility of memory - the visuals were frankly baffling, less about narrative, mood, tone, communication etc. than they were filler, just loads, and loads of pointless visual fat inside an already extremely short, short film.
Plus ok like he doesn’t want to kill himself because of the responsibility to the people whose bodies he inhabits, but he agrees to fly to Paris every day in a new body, no matter what, to meet up with Quinn? Isn’t that also irresponsible? Aren’t you also just completely fucking up these people you inhabit when you fly over, meet your partner, and then abandon these poor vessels to navigate this country alone? What do their families think? How many marriages did you plan to destroy? Did you bankrupt them? What happens after you pilot their bodies to Paris, you selfish fuck? This is a story about a solipsistic little weirdo who does not see the people around him as agent human beings. Not a single second is spent contemplating the fate of the prisoner he is trapped in, the only thing that matters in this tiny world is his own experience and nothing else. Fuck you, got mine: a love story.
But see, I did watch it. I did have something to say. I treated it like a real piece of filmmaking, like I said I would. And it was very, very bad.
While trawling through other AI film channels it was mostly fake trailers for fake films, make-believe verticals for products that will never exist. Not exactly breaking down the barriers to creative work here so far, but fair shakes to PJ who did bring something to assess.
Part 2: Big Studio Energy
I want to re-emphasise that I watched a lot of AI stuff that wasn’t worth mentioning at all because it was incomprehensibly bad. Like I really did take the task of trying to find output that could be evaluated as a piece of creative media seriously, and there is so little out there. Insane bible videos with 800k views; barely legible slideshows; unwatchable waste product from great factories of shit; thousands, thousands of videos that exist exclusively to promise something TBD vaguely in the future rather than to be something.
Water, water every where, nor any drop to drink.
That being said, this is a powerful new tool, SO I’VE HEARD, and I understand that people are setting up slightly suicidal* studios to take it seriously, so I thought it would be good to look at a few of them and see if they’ve used it to create something.
First up: Staircase Studios. Staircase was founded by one of the producers of Divergent, Pouya Shahbazian, who’s producing credits also include Love, Simon and the 2024 Russell Crowe crime thriller Sleeping Dogs. They’re aiming to put out a film called The Woman with the Red Hair in the summer of 2025 (WHOOPS - I get it though, I do the same #relatable) using cutting edge AI generated tools and performance capture. Seems promising!
When you click on their one and only video, you get treated to some behind the scenes footage talking about the staff and the processes before it gets to the real meat of the thing: the first few minutes of their feature film.
The first minutes are lazy. This is lazy. It’s just lazy work. I don’t know how much money they spent to produce this or to try and mash the captured performances and the AI performances together, but it doesn’t work at all. There’s nothing here, again, other than one shot that evoked a degree of inescapable and unknowable dread right at the start, which was… apt.
What a funny shot. I cropped it a little (again, for fun) but what a deeply unsettling way to begin.
There is so much going wrong that to start would be to never finish. If you want you should watch it yourself. I’ll not provide the link here because some priors are too strong to abandon in the pursuit of critique, but this is just a waste of your time. They have dressed up the fake vertical for the fake film as if it was real, and they’ve squandered the efforts of multiple people for lazy, deeply lazy, garbage. Next!
Actually, not next. Hollywood quality lads, clip it and ship it. We're 100% making that Summer 2025 deadline.
ACTUALLY, NOT NEXT. So in this scene the flag is supposed to be hanging from the building, showing a Nazi-occupied Netherlands. Why not zoom in to see if you can spot any of the clever visual effects used to make this come to life.
Promise is a company that has generated a lot of media hype around being an AI production studio centred around AI filmmaking led by Jamie Byrne of Youtube and George Strompolos of Fullscreen. Now, Fullscreen is a very interesting company, but let’s leave all our baggage by the wayside, we’re here for new media and the potential it brings.
On Promise’s website they have a trailer for a project titled NinjaPunk. It’s… fine. It’s honestly phenomenal as far as AI trailers go, they clearly used real people to do a lot of real stuff, which means they dodged most of the uncertain, changing ghost spaces that AI film production seems to take place in. I saw more than one shot take place in the same room and gasped out loud - finally, food that has successfully found the plate.
The glaring issue is that this is a trailer. It’s just a trailer. It is barely a proof-of-concept, and it’s throwing around shiny blades and rain-slicked roads as baubles, distractions from the cold, hard fact that it doesn’t look very good. I’m super sorry to the team, but all this effort in developing an AI-forwards pipeline has gone into creating something that is functionally and narratively identical to an advert for an AMD GPU.
Their accompanying BTS video shines a light on the process as an AI native studio (after a sizzle reel featuring zero projects advertised on their site lmao) and it actively looks like AI is getting in the fucking way. My favourite part for sure is when the voiceover by chief creative officer and third co-founder Dave Clark says “With GenAI we are able to ideate and iterate rapidly” before showing a series of generic images that only barely have any connection with the trailer. Truly this system is so powerful that GenAI is making me ideate right now! Glad they got more funding though, happy for them.
But again, I am struck by the sheer absurdity of it. Where’s the production. Where’s the intent. Where’s the story. What do you have to hook me into the world of NinjaPunk. Do you have anything to show me other than another trailer with these powerful creative tools you are using? Did you try to do anything other than provide a vector for more investment while like… pretty blatantly ripping off Ghostrunner?
Look at this image. With the power of AI, the team at Promise has made it move. Why aren't you clapping? Why aren't you giving me twenty million dollars in VC funding?
Dave Clark has his own Youtube channel for AI videos called @Afraid2Sleep (or more colloquially ‘AI Director Dave Clark’) which does have proof of concept short films, but isn’t linked or embedded anywhere on Promise’s website. Why?
If you are making AI films, breaking down the barriers with these powerful, democratising (so I’ve heard etc. etc. etc. we’ve run this particular joke dry) tools wouldn’t you want to highlight the storytelling that you can currently do with them, the stuff that you already have on hand that likely contributed to Dave becoming the CCO in the first place? I’m not going to give these a review (even though they are real, actual short films) because I feel like they’ve been shamefully hidden for the sake of trying to sell pure potential, but Dave uploaded a video as recently as eighteen hours ago (at time of writing obviously) so he clearly cares about it enough to put non-Promise time into the project.
I can speculate as to the possible reasons why Dave’s output is divorced from Promise’s promises, but this is already too fuckin long and I don’t want to engage in it anymore!
Ah, finally, a source for some of the AI power and influence claims! This information is confirmed TRUE by REAL AMERICAN PATRIOTS
Watching the rise of AI shit makes me feel like I am losing a degree of understanding that I thought I had with the people around me. It is draining my ability to empathise with others. When I watch this ‘content’ this ‘output’ I am reminded that there are people who will swear up and down that this is either the same as a crafted film, or better, and it throws me a little. It truly throws me. It makes me think that I don’t exist in the same world as them, like they hear different chords from me when they listen to music, like they see completely different images when they view a film, taste something completely different when they eat food. A world where there is no commonality of experience between us even at the most base level. Each day I encounter AI arguments for my job, and each day I worry that I feel less human - this stuff corrodes, it creates gaping holes in the hull of the ship and I feel myself dragged out into the inky void to view it all from a deeply unsafe and isolated distance.
I wonder sometimes - did you ever enjoy reading a piece of copy, ever? Even once? Did you ever take pride in something you wrote, approved, or otherwise released? An e-mail? An essay? Did you ever enjoy seeing an advert? Did you ever genuinely like a friend’s picture? Or a painting? A home movie? A television show? A film? I’m not sure I can believe in people anymore when so many are so willingly give up their relationship to creating anything for the sake of what - a theft robot? Grammarly? The world’s worst therapist?
How far from me do you have to be to watch something like PJ’s Lord of the Rings adaptation (generous phrasing) or The Woman with the Red Hair or NinjaPunk and go “man that looks good” - what a deeply, profoundly distressing gulf to be aware of.
*'Suicidal' is a bit strong, no?' Ok walk with me here - you surely understand that there's a very, very narrow window of time where film people can make money from this. Right now, even if you're positive about it, there's a knowledge gap that is only barely exploitable. An AI studio will naturally be in a completely tech-determinative race with other AI studios (and AI platforms) to see who can drain as much money from suckers as possible before another studio with another platform showcases something easier that makes your prompting expertise completely obsolete. It's suicidal, it is a death-drive to grab like fifteen extra bucks from VC freaks.
A brief note about sources: I have deliberately not included any links to any of the AI generated videos mentioned or the pages that host them. Sorry, lmao.
No comments:
Post a Comment