Could an individual using nothing more than neural networks combine imagination and production into a single process and single-handedly oversee the creation of an acclaimed and commercially successful movie? Is AI-generated art going to be used to stimulate new forms of expression that transcend what we see on our screens today, spawning new styles and genres? Or will studios just use it to create digital doubles of A-Lsters so films can be made by fewer people for less money and at lower risk?
Writing on Medium, Pinar Sehan Demirdag thinks we will see the creation of content using artificial intelligence arrive in the film industry in three stages. Visionary producers will use it for title sequences, which allow for more creative freedom. From there it will be introduced more into volume stages and ICVFX before making its way into concept art, storyboarding, animatics, pre-visualisation and post production.
After this will come offline VFX, both 2D and 3D, once neural networks have successfully nailed real-world physics, temporal cohesion and the language of filmmaking in general, and text-to-3D takes off.
Combined into a kind of virtual film generation studio, they will give computer-based control to light, coherent foreground and background generation, video-driven motion, performance, camera movement and the generation of depth maps. Please get in touch if you would like to speak to Mondatum about creating systems like this for your own needs, please get in touch – firstname.lastname@example.org
Main image: from Salt