ASK & DISCUSS INDEX

2 years, 5 months ago - Stefanos Christofi

This is all probably wrong so take it with a massive pinch of salt.

I think in the next few years it will be something like pretty much everybody can probably create most stuff to a certain extent e.g. make me an advert for jeans but it will be difficult to get exactly what you want especially the more complex and precise the requirement (because AI works on probability and doesn't have a clue really what you mean) is and that will be the crux of it.

I think AI will struggle to convey the correct emotions not least because they are hard to define exactly. AI might kill things like extras roles eventually but I think actors are safe at least for now. I think possibly locations will not be needed as much and effects and editing may be easier but again I think, precision is where AI struggles so to get truly great stuff will take knowledge and mastery probably of both the AI and creativity. Which actually brings on another issue, AI models appear at a silly rate where its difficult to keep up with the features let alone master them and by the time you do something else will probably have come out to supersede them.

So thats my sense AI will probably give everyone the ability to create quickly and cheaply but it will still take real skill to create great things especially great things that are popular too.

2 years, 5 months ago - Sarah Sawyer McCarthy

I'm fascinated by this concept at the moment. I've been asking CHATGPT to write monologues and have been performing them online. You can watch on IG @seesawmac

What I'm noticing at the moment is that the software creates very general concepts and non-specific work, so I have to often tailor the questions I ask it and give it lots of specific prompts. It also requires me to tailor it for the social media platform both in performance and how I edit it. Even when I put the video in an AI generated editing app, I still re-edit it afterwards. I do wonder what the relationship between creativity and AI will be moving forward. Right now, It's easy for AI to write a play based off a Sam Shepard play, for example, but coming up with something completely new and fresh-- the technology is no where near that yet and doesn't seem to be the primary focus of the developers. What I do think though, is that art is going to have to be more daring and out of the box moving forward. We will have to stay on top of the human experience, and we will definitely be using AI to aid our turnover rates esp. now that social media requires INSANE turn around times.

Long story short, we're okay for right now. But as the technology develops, we'll need to find ways to produce content that the technology cannot produce.

2 years, 5 months ago - Eric Garson

Dear Elliot,

Some general thoughts on your topic:

Concept Art: Recently an AI system created concept art of what Tron (1982) might have looked like if Alejandro Jodorowsky had made it. I believe Jean Giraud Möbius worked on the concept art for Tron, whom Jodorowsky worked with on Dune and The Incal, thus the imagery feels like it fits: https://www.nytimes.com/interactive/2023/01/13/opinion/jodorowsky-dune-ai-tron.html

Screenwriting: Maybe an AI system could scan and amalgamate all the screenwriting books ever written from Aristotle’s Poetics, Alexander Mackendrick’s On Film, Sidney Lumet’s Making Movies, Dan O’Bannon’s Guide to Screenplay Structure etc etc and create the ultimate screenplay. Not just in narrative terms, but also based on previous box office performance. Wouldn’t that take out a lot of the risk for productions/investors? But does that mean only films that are predicted to make money get made? And not art for art sake?

Spielberg: Spielberg recently said “if it ain’t on the page, it’s not on the stage.” https://youtu.be/hJhc66Ebb2c With that in mind, perhaps AI could enhance a screenplay, help create the stories backbone/structure and the writer can then add their human input later?

Corridor Crew: They recently made a short film called “Rock Paper Scissors” where they created an anime cartoon from a video using a program called Diffusion. I found the making of, more interesting then the film.

https://www.youtube.com/watch?v=GVT3WUa-48Y - Anime Rock Paper Scissors

https://www.youtube.com/watch?v=_9LX9HSQkWo - Did We Just Change Animation Forever?

Slightly separate topic, sometimes I see VFX/CGI attempting to look real and my sixth senses aren’t persuaded, deepfakes come to mind. What I found interesting is that because it was an anime cartoon it’s meant to look unreal/abstract, it worked, for me at least.

The technology is free/cheap and out there now, it’s not going away. Thus I think people will take advantage of it once they're familiar with it. My hope is that it’s a tool to help, my concern is that it might replace jobs.

Disney Animator: Furthermore there’s an interesting response by a Disney animator to the corridor video: https://youtu.be/xm7BwEsdVbQ

Funding: For so many filmmakers the main barrier is funding, I know this personally. Filmmaking is the most expensive art form. Digital filmmaking has opened up the craft to anyone who has a mobile phone to an Arri Alexa. But it still costs money to make a movie: cast/crew wages, locations fees, accommodation etc. If AI could match the right investor/organisation to the right filmmaker/film, help find funding or make a project more commercially viable that would be great.

Final: From what I’ve seen of recent developments in AI, virtual production and Unreal Engine et al. It appears the majority of projects made appear to be sci-fi heavy. It’s rare to see an 18th century period drama, comedy, musical, Biblical epic etc utilising this technology. Please correct me if I’m wrong, as I would be interested to see.

Last, I would love to meet someone in person who can explain Unreal Engine to me. Hopefully I have some writing/storyboarding chops (if thats not replaced by AI) but I don’t have the VFX chops. Thus I’d love to meet someone who has a expertise/passion in that area, especially if Berkshire/Oxfordshire/Buckinghamshire based.

2 years, 4 months ago - joe lee

Asked chat-gpt and they said:

Generative AI technology is rapidly advancing, and its applications in film making are becoming increasingly sophisticated. One significant development is the use of generative AI algorithms to create realistic computer-generated imagery (CGI) that can be seamlessly integrated into live-action footage, allowing filmmakers to bring their wildest visions to life on the screen.

Another area of development is the use of generative AI to automate certain aspects of the filmmaking process, such as editing and color grading. AI-powered tools can analyze footage and make suggestions for edits, allowing filmmakers to streamline their workflow and spend more time on creative tasks.

In addition to these practical applications, generative AI is also being used to inspire and facilitate creativity in filmmakers. For example, some AI systems can generate story prompts or plot outlines based on user input, helping filmmakers generate new ideas and explore different narrative possibilities.

Overall, the impact of generative AI technology on creative filmmakers is likely to be significant. By automating certain tasks and providing new tools for inspiration and experimentation, AI could help filmmakers create more innovative and exciting work. However, there are also concerns that increased reliance on AI could lead to a loss of human creativity and originality. Ultimately, the relationship between AI and creative filmmaking is likely to be complex and multifaceted, with both benefits and drawbacks.

2 years, 4 months ago - Lynwood Shiva Sawyer

Hi, Elliot,
If you want to know what impact AI will have on the creative (not the technical) process involved in making films, you only need to look at AI did to the music industry.
I don't have an office, so I work in coffee shops in the morning and pubs at night.
The barristas and bartenders are generally in their 20s to early 30s, and whoever is on deck gets to choose the playlist.
Inevitably, they pick music from the 1960s through 1980s.
I pointed that some of the songs were 50 years old. It would be like in 1960 listening to music from 1910!
"But the music was so much better then!" the barristas/bartenders, inevitably replied.
Around seven or eight years ago, a research company (I could be getting the stats wrong, but the conclusions are correct - if anyone has the proper cite, please add) picked 500 top 40 songs from the last 50 years and played the intros to 500 millennials. Obviously, the researchers didn't play every subject every intro.
As I recall, the most recognized song, probably because it's played at every wedding and appears in every romcom - "When a Man Loves a Woman" and number two was "Respect" by Aretha Franklin. The millennials only recognized TWO songs that had been recorded after the year 2000.
Sometime in the 1990s, a company created an algorithm that analyzed all the top hits for the preceding 50 years to determine commonalities. (Forget their name, but if you have the cite, please post).
Now every major label runs their songs through the algorithm, which will then instruct the record label to decrease the bass, repeat the chorus, whatever.
The net result is obvious to any one who has had to endure more than half an hour of best selling contemporary music and not hearing a single song that sticks in their memory.
Music that accountants and MBAs love, which real human beings find indistinguishable. Fortunately, London, unlike New York, has lovely and vibrant indie music scene.
You scan see the same trend in Save-the-Cat tickbox movies, which will continue to drive audiences away from the cinema by their increasing predictably but hopefully will allow breakthrough movies like EVERYTHING to flourish.

2 years, 4 months ago - GAYNOR O'FLYNN

Hi

Perhaps read my Linkedin posts on this...

https://www.linkedin.com/feed/update/urn:li:activity:7051826323412918272/

Or check out

https://www.youtube.com/watch?v=y7g-nRooZr8

2 years, 4 months ago - Richard Lipman

AI technologies can compose, produce and record music and thus in turn poses a number of important copyright questions, including around the copyright status of music created using AI, and also what licences are required when such technologies are trained by crunching data associated with existing songs and recordings.

More info online...