Renowned filmmaker James Cameron, celebrated for his work on blockbusters such as "Avatar," "Terminator," and "Titanic," has recently taken a bold step into the realm of artificial intelligence by joining the board of Stability.AI, a prominent entity in the Generative AI sector. This move comes as a surprise to some, considering Cameron's past portrayal of AI in his "Terminator" franchise, where Skynet, an artificial general intelligence, becomes a threat to humanity. However, it seems the director is now embracing the technology, which he views as a significant advancement in the field of cinematic visual effects.

Stability.AI, valued at nearly a billion dollars, was once headquartered in a modest location above a chicken shop in Notting Hill. The company is best known for its Stable Diffusion technology, a text-to-image tool capable of generating hyperrealistic images from textual prompts provided by users. With its sights now set on AI-generated video content, Cameron sees the convergence of generative AI and CGI as the "next wave" in visual effects, as stated in a press release from Stability.AI.

In the film industry, filmmakers enhance live-action footage with special effects (SFX) and visual effects (VFX). SFX are the physical effects used during filming, such as explosions, crashes, and prosthetics, while VFX are digital enhancements applied during post-production, including CGI, compositing, and motion capture rendering. Cameron believes that the integration of generative AI into VFX will revolutionize the industry.

The advent of Virtual Production has brought VFX techniques into the filming process itself. This method employs "game engines," technology originally developed for video games, to project dynamic, pre-produced virtual environments around actors on sophisticated LED walls. While the physical nature of SFX means AI's impact will be limited, VFX presents a ripe area for AI transformation.

I will be discussing the topic of deepfakes and AI in film at a public lecture titled "Deepfakes and AI in Film and Media: Seeing Is Not Believing," scheduled for October 30, 2024. Additionally, the Synthetic Media Research Network, a collective I co-lead, is delving into this subject. The network unites film creatives, academic researchers, and AI developers. In conversation with Christian Darkin, a VFX artist and Head of Creative AI for Deep Fusion Films, he envisions generative AI in VFX as offering an infinite array of post-production options. In the future, filming actors may be just the first step, with the ability to add backgrounds, alter camera angles, modify expressions, enhance emotional performances, change voices, costumes, and even faces at will.

A primary motivation for the film industry to incorporate AI into VFX is cost. Traditional VFX are expensive, with numerous technicians involved in their creation. Generative AI presents a more cost-effective solution for achieving high-quality visual effects, potentially without compromising quality. This could lead to job losses among VFX technicians, but there is an optimistic view that these skilled professionals will transition into new roles in emerging tech fields.

The ethical considerations of AI technology are a pressing concern. Media creatives now have access to a vast array of generative AI tools that enable the creation of images, text, voices, and music in novel ways. However, it is crucial to question whether these AI tools have been developed ethically. Each tool, from ChatGPT to Midjourney to Runway, is built on a foundation model trained on vast datasets, often sourced from the internet, to enhance their capabilities. This training process involves the use of "crawlers," bots that scour the web for material and download trillions of files.

This practice has led to legal disputes over copyright. Stability.ai is currently involved in a lawsuit in the UK courts, with Getty Images, a major repository of images and photographs, suing the company. A former Stability.ai executive, Ed Newton-Rex, resigned in November 2023 over concerns about the company's practice of scraping creative content for model training without compensation and under the guise of "fair use."

Cameron may believe that AI developers will prevail in these legal battles and continue their technological progress. When asked whether Stability.ai had used any of Cameron's creative material from the internet as training data for their foundation models without his permission, the company responded, "We're not able to comment on the source of Stability AI's training data."

Cameron's "Terminator" films cautioned against the potential dangers of unchecked AI, yet the director now appears to be betting on the success of AI in the film industry. As an Associate Professor in Filmmaking at the University of Reading, Dominic Lees explores these themes in his work, which is republished from The Conversation under a Creative Commons license.