AI Tools Bring Efficiencies But Have Limitations, Says Weta CTO Kimball Thurston: “Tech Companies Should Collaborate More With Creatives”
Weta FX chief technology officer Kimball Thurston said AI tools can make film production faster and less costly, while warning that technical and design limits remain. At Netflix’s Creative Asia event at the Busan International Film Festival, speakers argued that AI should expand filmmakers’ creative choices, with human judgement still central.
Thurston said Weta is using AI for image clean-up and generating crowd motion, while colour range and text-based interfaces remain problems that technology companies should address with creatives. Netflix’s Sung Q Lee cited pre-visualisation, editing and visual effects as uses, and said the company applies internal guidelines. Director Kang Yun-sung said AI helped cut costs on his upcoming film Gulf Of Aden by 30-40% by simplifying shooting and shortening production.
- Weta’s CTO says AI can speed up film production but still has technical limits.
- Thurston wants better colour handling and interfaces designed for visual artists.
- Director Kang says AI reduced costs on Gulf Of Aden by 30-40%.
New here? Start with this
Artificial intelligence tools are increasingly being used in film and television production, handling tasks such as cleaning up images, generating crowd scenes, and assisting with editing and planning. Major entertainment companies including Netflix and Weta FX are exploring these technologies to see how they can reduce costs and streamline production workflows.
However, AI tools currently have important limitations when used for creative work. Technical issues such as difficulties with colour accuracy and translating creative ideas into text-based instructions mean that human expertise and judgement remain essential to filmmaking.
This is significant because AI is reshaping the film industry. The technology is being adopted widely, but the industry is still working out how best to develop and use these tools to support filmmaking.