The latest generative models are great demos, but are they really going to change film and TV production? According to film producers and VFX experts, not in the short term, but in the long term the changes can literally exceed all imagination.
During a panel discussion at SIGGRAPH in Denver, Nikola Todorovic (Wonder Dynamics), Freddy Chavez Olmos (Boxel Studio) and Michael Black (MeshcapadeMax Planck Institute) talked about how generative AI and similar systems can change – but not necessarily improve – the way media is produced today. They agreed that although we can rightly question the usefulness of these tools in the near future, the pace of innovation is such that we must be prepared for radical changes at any time in the subsequent period.
One of the first topics was the impracticality of today’s video generators. Todorovic noted that “it’s a misconception that AI works with one click and you get a final VFX shot, when in fact that’s impossible. Maybe we’ll get there, but if there’s no editability, this black box won’t do much. What we’re seeing now is that UX is still being explored – some companies are starting to learn how to express 3D and filmmaking.”
Black pointed out that language is fundamentally unable to describe some of the most important aspects of visual creation. “Just like yoga or ballet, there are some classical things that have names that we can define, but most of the things we do don’t have names,” she said. “And there’s a good reason for that: because people have there is actually a behavioral model. But I don’t have a generative model of images in my head. If I want to explain to you what I see, I can’t project it from my eyeballs. So I have to use words and although we have many words to describe the visual world, but if I want to describe a certain movement to you, I will simply show it with my body, and then you will see it with your eyes, and you will be actively involved in the understanding. And I think this is a biological reason , a neuroscientific reason we don’t have words for all our movements.”
This may sound a little philosophical, but the point is that text-based image incentive systems are fundamentally limited in how they can be controlled. Even the hundreds of technical and artistic terms used every day during filming and post-production are insufficient.
Chavez Olmos pointed out that being Mexican, he had little opportunity to participate in the world of filmmaking, as all the money and expertise was concentrated in Los Angeles. However, he said that AI expertise (and the demand for it) has attracted wider interest. “I had to leave Mexico because I had no opportunities there; now I see that the same opportunity exists for those who remain, they no longer have to go over the border to do so.”
But Black worries that sudden access to such procedures could have unintended consequences in the short term. “You can give someone a powerful car, it’s not going to make them a Formula 1 driver, is it? It’s a bit like that now. People are talking about everybody’s going to make movies. Honestly, they’re going to be crap,” he said. “The democratization thing is exactly what Chavez Olmos said, and the power lies in the fact that maybe a new voice gets an opportunity that it wouldn’t get otherwise. But the number of people who make really good films will always remain low, in my opinion.”
“The real revolution,” he continued, “the real power of what we’re seeing in artificial intelligence is that we’re going to see a whole new genre of entertainment, but I don’t know exactly what it’s going to look like. I predict that video games, movies and real life will be something. The film industry is a passive story-telling: I’m just a passive receiver of entertainment. But in everyday life we talk about what we did on the weekend and so on. And it’s very active interactive storytelling.”
But before that happens, Chavez Olmos said he expects a more traditional adoption curve for AI-generated images and actors. “I think the reaction will be the same as when we saw the first ‘Final Fantasy’ movie or the ‘Polar Express’ — it won’t be completely finished, but people will start to accept those movies,” he said. “And instead of a completely It’s going to be a CG movie, it’s going to be a totally OUR movie, which I think we’re going to see at the end of the year. I think people are going to get over that, ‘Okay, it’s artificial intelligence,’ people are going to accept that.”
“The most important thing,” said Black, “and Pixar taught us this very clearly: it’s all about storytelling. It’s all about connecting with the characters. It’s about the heart. And if the film has heart, it doesn’t matter if the characters are artificial intelligence, I think people will enjoy the film,” he said. “That doesn’t mean they won’t want human actors. There’s a thrill in knowing that these are real people who are just like us, but so much better than us, to see one person at the top, that inspires us all, and I don’t think that’s going to go away.” he said.
Source: sg.hu
