Increased or annihilated? What if artificial intelligence invades Hollywood

Vanity AI, the tool developed by Marz to practice the so-called deaging/aging (in Italian, rejuvenation or aging) of actors, was used so far in 27 productions, including films and series also of great or very great success, such as Spider-Man: No Way Homeseason 4 of Stranger Things or The Umbrella Academy. It happened without more or less any of us viewers noticing, which is both a good and a bad thing. A risk and an opportunity.

And this was precisely what was discussed during the AI ​​goes to Hollywood panel, which we attended at CES in Las Vegas (here all the news) and precisely dedicated to what could happen now that artificial intelligences are everywhere, and therefore also in the world of cinema.

Help or a scary thing?

Marz is a Canadian special effects company (the name stands for Monsters Aliens Robots Zombies) and its chief operating officer, Matt Panousis, said that the company was born in 2018 for the world of television, with the aim of reducing the costs of special effects: “With our AI we can do the deaging 300 times faster than before, meeting two important needs in the world of showbusiness”, namely “maintaining high quality standards and cutting times and costs”. Even avoiding the actor having to undergo several hours of makeup.

It happens virtually without the viewer being aware of it, but somehow we are the watchers who wanted it to happen: “Over time, people’s requests and expectations have also grown for productions intended for TV, so much so that we expect a quality in the special effects that is like that of the cinema. And with these tools, we’re almost there”. But with significantly smaller budgets.

Reduced costs are not the only advantage: “The use of AI has allowed ad actors with disabilities to overcome them in some way and to be able to act in films where maybe they could not have acted, of work remotelyto rejuvenate and also to get paid forever”, recalled Duncan Crabtree-Ireland, executive director of SAG-Aftra (one of the most important actors’ unions), referring to the case of James Earl Jones, historic voice actor of Darth Vader, whose voice was recreated by the Ukrainian startup Respeecher. Again: still with regard to dubbing, the use of AI allows a better synchronization of the movement of the lips to adapt to the different languages ​​spoken, so that a film seems to have been shot in Italian even if it is not shot in Italian, and also to scenes without actually reshooting them, reconstructing any changes in framingreplacing the face of an actress who is no longer part of the cast or redressing an actor who was initially undressed, as explained Scott Mann, co-founder of Flawless, whose AI is called TrueSync and specializes in dialogue.

Make a full movie on your home computer

These are all examples of generative AI, i.e. not only analyzing or collecting data but also creating content from scratch, such as photos, texts, three-dimensional images, whose use “has been growing exponentially for 6 months now”, as the writer Nina Schick recalled, specialized precisely in GenAI. Who during the Las Vegas meeting made two rather daring predictions, saying that “by 2025, 90% of video content in the world will be made using AI” and that “within 5 years it will be possible make an entire movie on any computer that we can buy in the shop”.

Schick’s idea is that “creators will be able to do incredible things with the help of AI” and that when asked if we will be “augmented or automated”, which he did to the audience from the stage, the answer can only be the first: “Surely some works will be lost and will be automated, but others will be created and above all possibilities will be created that we did not have before”. What do you mean? “In the sense that writer’s block will no longer exist, because even if I lack inspiration, I could still turn to (for example) ChatGPT to get a trace, an idea, a starting point, a draft”. Or even a list of questions to ask for an interview.

Returning to the world of cinema, according to Schick “we are only at the beginning” e progressively “we will see a sort of democratization of creativity and the barriers to entry will be lowered, because it will be possible to have scripts, stories and special effects at a fraction of the current cost but maintaining a good level of quality”.

Cinema

3D is not the most amazing technology in Avatar 2 – what is the HFR system and how does it work

by Dario d’Elia

December 17, 2022

The risks of AI in the world of cinema

The key to quality is a sore point, because one of the fears of the use (better: abuse) of artificial intelligence in Hollywood is that we arrive at a kind of “acceptable mediocrity”, that is, that the quality of everything is lowered and that we who see and they who produce will make it go well for reasons of cost and comfort. An example to understand, made on the stage of the CES: “If a student has ChatGPT write the report on a book that he had to read but has not read, he will hardly obtain a text that will allow him to get a 9 or a 10 but very easily he’ll take home a 6. And he’ll be fine with it”. Here: will we be fine like this?

At the moment it is still too early to answer, just as it is too early to answer another of the questions related to this issue: what will prevent (for example) Disney from creating the next villain of the Star Wars films from scratch without even using a real actor or actress, just as virtual influencers are created? According to the panellists, the risk would be averted by empathy (or lack thereof) with viewers, which people are better at creating than machines. At least for now.

The last point is perhaps the most important, and it concerns copyright and its infringement: these AIs generate content but they don’t generate truly original content, because to obtain it they always start from the ideas of some human. In the case of a hypothetical screenplay for a film made by artificial intelligence, where does the idea really come from? Where do the databases on which the AI ​​has trained and learned come from? Where was the scraping done? Again: in the case of non-existent faces in the real world and made by a machine to interpret a film, whose original faces are they? Have the people to whom they belong authorized their use? Were they paid for it?

They are all sensitive and important questions, to which we should already be able to give an answer by 2023. We’ll make it? Yes, according to the participants of the Las Vegas panel: “Fear is understandable, but it must not take over – Crabtree-Ireland underlined at the end of the meeting – Even Photoshop was terrifying at firstand today we all use it to make beautiful things”.

@capoema