Is AI the future of film making?

[Matias Nicolas]

I feel that AI is gaining more ground in advertising productions. Even though many images still feel fake, it has the ability to create visuals with very high production value — the kind that would cost a lot of money to make in real life.

Do you think this will affect our work in the short, medium, or long term?

[Aristeidis Tyropolis]

It will affect everything and everyone because the rule of "good enough" will apply.  But we should be expecting some sort of "market correction" as there has been literally trillions spent (and still being spent) on the topic.

Models that cater to specific things (such for audio cleanup) are showing good benefits, same for image up-rez (here the results are more mixed)

Narrative work should be mostly unaffected - for a while.

In the end we (as a society) need to pick what we expect from this - so until this is answered it's really difficult to say where and with what we'll end up.

A significant portion of the cinéphile and even mainstream film maker crowd  is now using film again - for photography as well, so there's clearly hope, some still believe in the "tangibility" of things.

[Matias Nicolas]

I feel like one day we'll even abandon the use of optics as we know them and start using digital lenses that emulate traditional optics! Even though it seems impossible right now, I believe it will happen someday. It's as if everything ends up becoming democratized by lowering the cost of equipment. I've heard that Arri is in crisis because its lights are too expensive, because more affordable LED lights from other brands have appeared.
And this AI thing seems here to stay and is advancing very quickly. Production companies are already developing a pool of advertising directors who use only AI.

[Jon O'Brien]

Make the films you want to make if you can make them, and my advice (not that I get listened to much by anyone in the film industry) is shoot on film, and state boldly in the credits that "This film was shot entirely on real photographic motion picture film, and no AI or other digital input or manipulation occurred in its production." George Lucas said once that all he did was make the films he wanted to make, to please himself. Aim for that, even if you're making a 2 minute short. As for how much AI will affect jobs, I'd be starting to become a bit concerned if I earned my crust from anything creative that is 'digital.'

Edited October 15 by Jon O'Brien

[Aristeidis Tyropolis]

The flip side is that a lot of people will look for authenticity, and generic image/video LLM's like Sora are the epitomy of the statistical average.

Advertising was always an industry of "averages" and cookie-cutter faux lifestyle-selling borne out of the cannibalisation of counter-culture. 

Authentic approaches to story telling and visual approach are very much needed, look how successful PTA's new movie is - clearly there's opportunity there.

[Matias Nicolas]

It's all very clear, but beyond films and authenticity, will AI replace a large part of commercial production? Or will more possibilities open up? Because the creative approach one usually seeks is clear, but since the advent of digital, plus social media, and YouTube tutorials, it's harder to differentiate yourself from the rest. I'm talking about this in general. When I started working as a camera assistant in film, there was film, and then there was everything else. Today, it's more like a logarithmic curve, where now there are many new people in the middle who find it easier to film and get good results. Added to this is the reduction in equipment costs. On the one hand, you have the artificiality of AI, narrative advertising like the one we've always been doing, and I feel that authenticity is sought more in influencers showing products in their homes.
I don't know if this will even reach series and movies in the future!

[Aristeidis Tyropolis]

My personal opinion is that it will. If I was a camera operator at this time juncture, I would primarily focus on narrative and documentary (also tons more interesting)- maybe supplement income by working with some YouTuber properly showcase material - would stay the heck away from weddings.

[Tyler Purcell]

I was reading a few studies on this and it's very fascinating. 

Younger audiences are ok with AI stuff, they seem to not be really upset. It's the middle age and older people who don't like it. 

This is a bad sign, because I had always assumed younger people who be the ones who freak out because so many are tired of fake shit, but it seems like I was wrong. Seems like they've embraced it to a point where they know it's fake but are ok with it. They are using the tools like crazy to be creative without spending any money. 

With that said, there is no way vendors are amortizing the cost of high quality AI image generation, zero. The entire business model is a net negative, with some estimates being over 200 billion dollars spent thus far and over 2 billion being LOST per month. The whole thing is a circle jerk, with hardware companies investing in AI companies and AI companies investing in hardware companies. The whole thing is being propped up by the stock market, similar to the dot com and housing bubbles which led to the last recessions. So every analyst says the entire business is going to collapse, maybe not this year, but sometime in the very near future. Once it collapses, the dust from the disaster will force them to raise prices substantially, to the tune of several thousand dollars a month in order to stay healthy financially with automated video creation especially. I don't think ChatGPT will have a huge increase. 

Local LLM's aren't good enough for high quality video generation. So there is no way to put it into a camera, at least not anytime in the coming years. We are slowly starting to see some still image regeneration, programs like Pixelmator Pro has a really neat tool which pulls from the still image you're currently using for patching. It won't pull from a database tho, only the single image you're using. It's the database pulling which makes the video AI tools so powerful and I don't see how any of that will be made local, maybe even ever. The only way for generative video AI to be successful, is to charge a lot of money for it, period. Thus, it will be pulled away from consumers shortly, maybe within the next year once things start to collapse for them. 

Unfortunately the damage has already been done, ad agencies can pay for instant results rather than hiring a crew and maybe not getting what the client wants. As the tools raise in cost, they will also get better. I predict in the next 2 years, we will see user manipulation become common place, where operators can drag things around on layers within a generated scene, which will allow FAR more customization. Agencies will pay for that, they will also pay for that single person who knows how to use the machine, but the days of crews on set? Long behind us, especially for content going on the small screen. Eventually all advertising will be done with AI, because clients will be attracted to the instant results and abilities to change things later, which is hard to predict when on set. Having worked in commercial production since the 90's, I can tell you it's DEAD in the United States. It's so dead, I'm shocked there isn't a tombstone at the Hollywood Forever cemetery. LOL 

Edited October 15 by Tyler Purcell

[Joerg Polzfusz]

There’s no AI yet. There are only LLMs that can repeat and remix things that they have „learned“. In two years, the LLMs might be good enough for everything that needs to be cheap. But their Achilles heel is that the LLMs aren’t creative: They can probably „rewrite“ Shakespeare in the style of Quentin Tarantino and then produce a video in the visual style of „Frozen“. But they can only do this when having been trained with Shakespeare, Tarantino and Frozen and only when a human asks them to do this. But that’s it. They cannot create anything that is completely new. So they are only copycats (or maybe even copycat criminals, depending on what they have been trained with - https://www.npr.org/2025/06/12/nx-s1-5431684/ai-disney-universal-midjourney-copyright-infringement-lawsuit ).

[Karim D. Ghantous]

15 hours ago, Matias Nicolas said:

I feel like one day we'll even abandon the use of optics as we know them and start using digital lenses that emulate traditional optics! Even though it seems impossible right now, I believe it will happen someday.

What does that mean, exactly? Like, fibre optics? Perhaps there is a better word you want to use than 'digital'.

As for advertising photography (not TV commercials), things are great for the more competent photographers and studios:

(9:06)

[Jon O'Brien]

45 minutes ago, Karim D. Ghantous said:

 

(9:06)

I like the guy's attitude in this video and I agree with him. Do better than AI can do and people will seek you out.

 

[Matias Nicolas]

16 hours ago, Tyler Purcell said:

I was reading a few studies on this and it's very fascinating. 

Younger audiences are ok with AI stuff, they seem to not be really upset. It's the middle age and older people who don't like it. 

This is a bad sign, because I had always assumed younger people who be the ones who freak out because so many are tired of fake shit, but it seems like I was wrong. Seems like they've embraced it to a point where they know it's fake but are ok with it. They are using the tools like crazy to be creative without spending any money. 

With that said, there is no way vendors are amortizing the cost of high quality AI image generation, zero. The entire business model is a net negative, with some estimates being over 200 billion dollars spent thus far and over 2 billion being LOST per month. The whole thing is a circle jerk, with hardware companies investing in AI companies and AI companies investing in hardware companies. The whole thing is being propped up by the stock market, similar to the dot com and housing bubbles which led to the last recessions. So every analyst says the entire business is going to collapse, maybe not this year, but sometime in the very near future. Once it collapses, the dust from the disaster will force them to raise prices substantially, to the tune of several thousand dollars a month in order to stay healthy financially with automated video creation especially. I don't think ChatGPT will have a huge increase. 

Local LLM's aren't good enough for high quality video generation. So there is no way to put it into a camera, at least not anytime in the coming years. We are slowly starting to see some still image regeneration, programs like Pixelmator Pro has a really neat tool which pulls from the still image you're currently using for patching. It won't pull from a database tho, only the single image you're using. It's the database pulling which makes the video AI tools so powerful and I don't see how any of that will be made local, maybe even ever. The only way for generative video AI to be successful, is to charge a lot of money for it, period. Thus, it will be pulled away from consumers shortly, maybe within the next year once things start to collapse for them. 

Unfortunately the damage has already been done, ad agencies can pay for instant results rather than hiring a crew and maybe not getting what the client wants. As the tools raise in cost, they will also get better. I predict in the next 2 years, we will see user manipulation become common place, where operators can drag things around on layers within a generated scene, which will allow FAR more customization. Agencies will pay for that, they will also pay for that single person who knows how to use the machine, but the days of crews on set? Long behind us, especially for content going on the small screen. Eventually all advertising will be done with AI, because clients will be attracted to the instant results and abilities to change things later, which is hard to predict when on set. Having worked in commercial production since the 90's, I can tell you it's DEAD in the United States. It's so dead, I'm shocked there isn't a tombstone at the Hollywood Forever cemetery. LOL 

Really interesting points. I agree that younger audiences embrace AI naturally!

My son told me some weeks ago:  Dad, I don´t watch commercials, I only watch influencers in tik tok!

I hope human creativity will still be key and AI will handle repetitive tasks, while people focus on storytelling and direction. For us, it’s another big transition, similar to when we moved from film to digital — a tough shift for many of us and for countless companies. Overall, it’s a challenge, but also full of opportunities for those who adapt... But 

[Matias Nicolas]

13 hours ago, Joerg Polzfusz said:

There’s no AI yet. There are only LLMs that can repeat and remix things that they have „learned“. In two years, the LLMs might be good enough for everything that needs to be cheap. But their Achilles heel is that the LLMs aren’t creative: They can probably „rewrite“ Shakespeare in the style of Quentin Tarantino and then produce a video in the visual style of „Frozen“. But they can only do this when having been trained with Shakespeare, Tarantino and Frozen and only when a human asks them to do this. But that’s it. They cannot create anything that is completely new. So they are only copycats (or maybe even copycat criminals, depending on what they have been trained with - https://www.npr.org/2025/06/12/nx-s1-5431684/ai-disney-universal-midjourney-copyright-infringement-lawsuit ).

Sure, but how do we fit into this game, when we focus on using cameras to film and we’re not screenwriters or directors? Is AI really going to do our job, or will we be the ones who evolve alongside the tool?

[Matias Nicolas]

8 hours ago, Karim D. Ghantous said:

What does that mean, exactly? Like, fibre optics? Perhaps there is a better word you want to use than 'digital'.

As for advertising photography (not TV commercials), things are great for the more competent photographers and studios:

(9:06)

What I mean is, just like we transitioned from film to digital, maybe in the future they could create cameras with some kind of digital optics — where, just like you apply LUTs in-camera, you could apply a specific lens brand or a certain type of flare digitally, and it would look so real that it seems like you’re actually using that brand of lenses.

[Matias Nicolas]

7 hours ago, Jon O'Brien said:

I like the guy's attitude in this video and I agree with him. Do better than AI can do and people will seek you out.

 

Great video! Love it!

[Joerg Polzfusz]

As a cinematographer, I would rather be afraid of being completely replaced by the camera:

• Your smartphone already has a follow focus that can properly focus the cat that is moving around in the image.
• Your smartphone can detect whether the heads in the image are pointing to the camera or not, whether the persons in the image are smiling, whether the person in the image is your aunt Bettie or your uncle John, …
• Veo, wan, … show that the computer can „understand“ what is a close-up, what is a medium close-up,…
• …

So simply put this logic into a computer that can control one or more camera drones (at least two for shot-reverse-shot) and feed it with the script and the storyboard - and the camera drone(s) will position themselves in the correct positions, will focus and zoom correctly and will wait for the director (speech recognition) and the fully autonomous audio- and clapperboard-drones…

(I guess that something like this would already be possible today - except for the drone’s power consumption and its noise.)

Edited October 16 by Joerg Polzfusz

[Mark Dunn]

Fortunately, AI can't service a Steenbeck and never will.

Edited October 16 by Mark Dunn

[Karim D. Ghantous]

16 hours ago, Matias Nicolas said:

What I mean is, just like we transitioned from film to digital, maybe in the future they could create cameras with some kind of digital optics — where, just like you apply LUTs in-camera, you could apply a specific lens brand or a certain type of flare digitally, and it would look so real that it seems like you’re actually using that brand of lenses.

So, camera software that processes images in real time to emulate certain lenses? I can't see how the lens itself would play a part in this.

[Uli Meyer]

In the near future longform storytelling could become a niche thing. People abandon cinema and home screens and instead get their entertainment fix doomscrolling through AI generated slop and mini soaps. That's already happening as attention spans decrease further and further. Is it inevitable? Who knows for sure. A friend who is working for a wealthy Chines producer just told me that his bosses have just abandoned their plans for feature film production and are investing heavily in vertical AI entertainment. There certainly seems to be a goldrush mentality which also partially explains why traditional production and distribution is in decline.

 

[Jon O'Brien]

Yes, well something weird is going on.

Very few interesting movies at the cinema. On television, unending repeats of Antiques Roadshow, I Dream of Jeannie, movies made about 15 or more years ago, and trashy boring movies on SBS world movies. Nothing Australian that I can remember. We survive on DVDs.

Can't ever remember such a flat period in Film & TV.

It's a dead calm. Mediocre look of films now. Flat looking. Not much talent.

Screen organisations shifting focus to gaming. An amazing amount of short film festivals being promoted left right and centre. It seems to be a quiet desperation by governments to help fund and encourage story telling.

Where are the big talents? The Peter Weirs? The Bruce Beresfords? The Kennedy Millers?

Everyone waffling on about AI.

It's a fad. People will be sick of AI movies. Doesn't matter how good they get.

Edited October 17 by Jon O'Brien

[Jon O'Brien]

Digital cinematography is a big part of the reason that the film industry went downhill. It all looks and feels like video, folks. Cheap look! Not good enough. Don't care what any big name cinematographers say about the topic.

Edited October 17 by Jon O'Brien

[Tyler Purcell]

5 hours ago, Jon O'Brien said:

Very few interesting movies at the cinema. On television, unending repeats of Antiques Roadshow, I Dream of Jeannie, movies made about 15 or more years ago, and trashy boring movies on SBS world movies. Nothing Australian that I can remember. We survive on DVDs.

This is because everything needs to be "safe" today. Marketing is too expensive. Production is too expensive. So if the studio's/streamers are going to invest in anything, it needs to be squeaky clean and made by a committee. Some good content slips through the cracks, but majority of it isn't marketed at all, so consumers don't watch it. Also, most of the good shows, tend to fizzle out because they stick to their formula too closely. It's almost as if TV/streaming shows are designed for audiences to fall asleep watching. 

5 hours ago, Jon O'Brien said:

Can't ever remember such a flat period in Film & TV.

Never in my life nope, this is a new problem. 

I recall going to the movies every single week when I use to be single. I saw literally everything, but once film projection went away, it seems the quality of the films did too. I recall 2014 being the first really dead year, with just a hand full of movies that were cool, from only the top creatives left. Since then, it's been getting drier and drier. 

5 hours ago, Jon O'Brien said:

It's a dead calm. Mediocre look of films now. Flat looking. Not much talent.

Everything looks the same because it's all done out of fear. They all use the same tricks, same toolkits, same volumetric sets, with noticeable CG everywhere even in basic dramas. 

"Oh that city doesn't look quite like what we're after, so lets make it all CG" LOL 

I was watching the new Superman movie recently, what an effin' bore. I don't think there was a single shot in the entire movie which wasn't heavily visual effects. Even at his farm visiting his parents, it was full of fake lens flairs and such. It's like, because they "can" manipulate, they "must" manipulate to make every single frame artificially pretty. 

We watched the re-release of Seven in IMAX last year with one of my close filmmaking friends who didn't know the film well and he was blown away. Not only does it look like a movie, but holy crap was the restoration good, looked like it was shot yesterday. We both agreed, that's what movies should look like. 

5 hours ago, Jon O'Brien said:

Screen organisations shifting focus to gaming. An amazing amount of short film festivals being promoted left right and centre. It seems to be a quiet desperation by governments to help fund and encourage story telling.

I do think in the last decade, the gaming business has taken off because it's a cheap form of entertainment. 

So many kids are lost these days and they have nowhere to turn, so social media, gaming/VR are basically their lives. So the gaming industry is very much on a roll. I don't see that changing anytime soon, but social media within the sub 24 y/o group is actually dwindling. They're living more and more on VR and with the advent of lower cost VR headsets, you can have an entire life on there that nobody knows about in reality. Ready Player One's reality is coming fast, you should watch that movie or read the book, that IS our future unfortunately. The "cinema" is pretty much dead once boomers and generation X start dying off. Their kids don't much care for the movies, they'd rather stay at home and play games or watch stuff in the comfort of their home. 

5 hours ago, Jon O'Brien said:

Where are the big talents? The Peter Weirs? The Bruce Beresfords? The Kennedy Millers?

There are plenty of them, but not from Australia.

I mean One Battle After Another is a pretty good representation of what American cinema CAN do. 
 

5 hours ago, Jon O'Brien said:

Everyone waffling on about AI.

It's a fad. People will be sick of AI movies. Doesn't matter how good they get.

AI is 100% a fad. 

[Jon O'Brien]

On 10/18/2025 at 5:30 AM, Tyler Purcell said:

... Marketing is too expensive. Production is too expensive ... once film projection went away, it seems the quality of the films did too. I recall 2014 being the first really dead year, with just a hand full of movies that were cool, from only the top creatives left. Since then, it's been getting drier and drier. 

I totally understand concerns about spending money, but whenever I raise the topic of increasing quality in Australian movies by shooting on film I get told by producers that "It costs a lot."

Yes, that's right. It does cost a lot. That's life. That's the movie biz. Want to make good movies that people actually want to watch?

What I find the most surprising is that so many people think the answer is to lower costs. No. The answer is to increase the quality.

As soon as producers (and for all I know investors too) started to see the savings that could come from digital cinematography and projection the film industry started rapidly to die. That's a fact. That's history.

Oddly perhaps, TV production followed cinema's lead.

Opportunities opened up with the changeover that the digital revolution brought and production quality declined as medicrity took over the industry. Tech heads, not artist types. Very, very obvious.

I think that shooting on film and projecting digitally is a formula that works for top quality production and screening. If you can project with film too, in select cinemas, all the better.

Edited October 19 by Jon O'Brien

[Edith blazek]

 

On 10/15/2025 at 3:59 PM, Tyler Purcell said:

I was reading a few studies on this and it's very fascinating. 

Younger audiences are ok with AI stuff, they seem to not be really upset. It's the middle age and older people who don't like it. 

This is a bad sign, because I had always assumed younger people who be the ones who freak out because so many are tired of fake shit, but it seems like I was wrong. Seems like they've embraced it to a point where they know it's fake but are ok with it. They are using the tools like crazy to be creative without spending any money. 

With that said, there is no way vendors are amortizing the cost of high quality AI image generation, zero. The entire business model is a net negative, with some estimates being over 200 billion dollars spent thus far and over 2 billion being LOST per month. The whole thing is a circle jerk, with hardware companies investing in AI companies and AI companies investing in hardware companies. The whole thing is being propped up by the stock market, similar to the dot com and housing bubbles which led to the last recessions. So every analyst says the entire business is going to collapse, maybe not this year, but sometime in the very near future. Once it collapses, the dust from the disaster will force them to raise prices substantially, to the tune of several thousand dollars a month in order to stay healthy financially with automated video creation especially. I don't think ChatGPT will have a huge increase. 

Local LLM's aren't good enough for high quality video generation. So there is no way to put it into a camera, at least not anytime in the coming years. We are slowly starting to see some still image regeneration, programs like Pixelmator Pro has a really neat tool which pulls from the still image you're currently using for patching. It won't pull from a database tho, only the single image you're using. It's the database pulling which makes the video AI tools so powerful and I don't see how any of that will be made local, maybe even ever. The only way for generative video AI to be successful, is to charge a lot of money for it, period. Thus, it will be pulled away from consumers shortly, maybe within the next year once things start to collapse for them. 

Unfortunately the damage has already been done, ad agencies can pay for instant results rather than hiring a crew and maybe not getting what the client wants. As the tools raise in cost, they will also get better. I predict in the next 2 years, we will see user manipulation become common place, where operators can drag things around on layers within a generated scene, which will allow FAR more customization. Agencies will pay for that, they will also pay for that single person who knows how to use the machine, but the days of crews on set? Long behind us, especially for content going on the small screen. Eventually all advertising will be done with AI, because clients will be attracted to the instant results and abilities to change things later, which is hard to predict when on set. Having worked in commercial production since the 90's, I can tell you it's DEAD in the United States. It's so dead, I'm shocked there isn't a tombstone at the Hollywood Forever cemetery. LOL 

But are people okay enough with paying for content from big studios created with it? Also a lot of the approval more comes from naivete as even among undergrads, the more aware people are about it, the less they are enamoured with it. Which for something that's hail mary for future survival is further saturation, is not good when the train is looking to crash any second now and make the dot com bubble and housing market crash look minimal in comparison.

[Edith blazek]

On 10/17/2025 at 9:33 AM, Jon O'Brien said:

Digital cinematography is a big part of the reason that the film industry went downhill. It all looks and feels like video, folks. Cheap look! Not good enough. Don't care what any big name cinematographers say about the topic.

Except, a lot of analog cinematography arguably looks not too dissimilar to digital, anyways, which would point to the culprit being their common factor, that being color grading.