The core of the argument is this: film is a language, and direction is the act of speaking it deliberately. Strip that out and there’s nothing left, no matter how polished the output looks.

Every choice in a film communicates something. Where the camera sits, how long a shot lingers, when a cut interrupts a beat, what’s in focus versus what’s blurred, whether the score swells or stays silent, which line in a scene earns the close-up. None of these are aesthetic flourishes; they’re sentences. A trained filmmaker watches a scene and reads it the way you read this paragraph, and when they direct, they’re choosing every word.

An AI tool doesn’t know what you’re trying to say. It produces statistically plausible footage: the cinematic average of everything it’s seen. That isn’t a film, it’s a series of generic gestures wrapped in reasonable lighting. And critically, the person commissioning it can’t tell the difference, because they don’t have the vocabulary either. They can’t notice that the eyeline is wrong, that a cut breaks rhythm for no reason, that the music is telegraphing what should have been subtext, that the geography of the scene is incoherent. They see “a film” and miss the dozens of small failures inside it.

This watches like you don’t have any experience making films, and if that’s true this is a communication failure at the source. You, as the filmmaker, can’t say what you want, can’t recognise it when you see it, and can’t iterate towards it. The AI fills this void with cliché, because cliché is the average. The result is a thing that looks like a film and contains nothing.

A useful analogy: you can use a generative tool to produce paragraphs of writing, but a person who has never read books, who doesn’t understand argument structure or voice, will produce noise even with the best tool. Not because the tool is bad, but because they don’t know what good would look like, so they can’t steer towards it. Film is harder still, because the language is non-verbal and the failure modes are subtler. A bad sentence reads as bad. A bad cut just reads as “a cut.”

Direction is taste in motion. Without years of watching, analysing, and trying to make things, that taste doesn’t exist.

Now, apply that to what’s actually on screen. The following elements show that this isn’t a film, it’s a two minute Instagram reel dressed up to go to the Oscars.

Vertical orientation. Cinema is horizontal because human vision is horizontal, and because the width of the frame is itself a storytelling tool. A wide frame lets you place characters in relation to each other and to a world. It lets the environment carry meaning. Vertical strips that out. You can fit one face, maybe two stacked, and nothing else. There’s no geography, no context, no sense of where anyone is. Vertical is the format of TikTok and Reels, designed for a single subject filling a phone screen during a feed scroll. It isn’t a cinematic frame; it’s a content frame. The choice signals at the outset that the maker is thinking in posts, not scenes.

Everyone facing the camera. Direct address to camera is, in film, a rare and deliberate device. Ferris Bueller does it. Fleabag does it. It works because the rest of the film maintains the convention of an unobserved world, and the break is meaningful. When every shot is a person looking down the lens, the device collapses. There’s no world being interrupted, because no world has been built. There’s no shot/reverse-shot, no eyeline match, no two characters occupying the same space and reacting to each other. It’s a series of talking heads strung together. That’s the grammar of testimonial videos, vlogs, and AI avatar tools, none of which are films, because none of them construct a scene. A scene requires people existing in relation to one another. This is portraits in sequence.

Floating text around each character. On-screen captions are grammar for the social media feed, designed for muted autoplay where you need to convey information without sound. Their presence in something claiming to be a film is itself the argument: if the visual and the performance were doing the storytelling, the words wouldn’t be needed. Cinema works through what isn’t said, through what’s implied by behaviour, framing, and silence. Captions short-circuit all of that. They also tell you something specific about the production. Generative pipelines struggle with coherent lip-sync and natural dialogue delivery, and floating text papers over both. People notice.

AI voiceover. Voice is performance, and performance is where character actually lives. The breath before a line, a syllable held a fraction too long, the catch in a voice that wasn’t in the script: these are what make a line mean something specific rather than something general. AI voiceover produces the mean reading of every sentence, technically correct, emotionally null, with no relationship to what’s happening on screen, because there’s nothing happening on screen for it to react to. Even if the maker wanted a particular reading, they couldn’t get it; and if they got it, they wouldn’t recognise the difference. So every line lands at the same temperature, and the result sounds like every other piece of content built with the same tools.

Stack those four together and you don’t have a movie with some rough edges and technical limitations. You have a moving video with completely different grammar: a vertical, frontal, captioned, voice-over format that is excellent at being scrollable feed content and structurally incapable of being a film. The form has already decided what the thing is, before the question of “story” even enters the picture.

In short: this will never be a good film. You might be able to Ship Of Theseus it into a good film, but only by learning how to direct and write, and then changing everything about it.

Even the absolute basics are incredibly bad. (black text on black background, shit kerning, the morphing car)

I have no words. Congratulations on… your other hobbies if you have any.

Hey dont let this get you down.

Lemmy hates AI. And with good reason. But these downvotes aren’t a reflection of your creativity or effort. I’m certain no one who has voted has watched yet. I haven’t, and probably won’t. But really, as someone who makes art, I know putting yourself out there and getting nothing but negativity can be really shitty.

I would search for an AI related community on Lemmy or off Lemmy to share AI art if you want honest feedback.

Good luck! :)

Edit: jfc this is an ai community. Sorry folks are being shitty. I don’t think much art is shared here.