At the Camden International Film Festival, a capital for documentary filmmaking, a group of documentary producers presented its best practice guidelines for using Generative AI in non-fiction filmmaking, also announcing further educational efforts.
The Archival Producers Alliance (APA) — founded in 2023 and made up of over 300 researchers, producers, and award-winning filmmakers — has outlined industry guardrails for the ethical use of AI as it relates to the usage of archival materials, like historical photographs to video footage, commonly used in non-fiction filmmaking.
The group has previously presented its guidelines in a working format, soliciting feedback from fellow documentary filmmakers, but is now presenting its final recommendations. Among other points, the best practices include a push for the use of primary source materials, but, when AI is used, considerations for algorithmic bias and usages of a watermark or other clear disclosures for audiences when an image was generated using AI.
In addition to releasing the guidelines, the group announced a multi-year educational initiative that includes a series of talks, panels and workshops across the country with the hope of helping documentary filmmakers ethically navigate using new technologies. The initiative, which will also conduct real-world case studies of using generative AI in the doc field, was launched with funding provided by the Jonathan Logan Family Foundation.
[...]
The APA’s guidelines have been endorsed by individual filmmakers like Burns, as well as organizations like the International Documentary Association, Documentary Producers Alliance, and Alliance of Documentary Editors.
“As documentary filmmakers who depend on the integrity of historical archives, and as mentors to younger filmmakers grappling with the threat – and the potential – that generative A.I. presents to their work, we are grateful for these guidelines,” said Firelight Films’ Stanley Nelson and Marcia Smith (Miles Davis: Birth of the Cool). “We are especially heartened by the attention given to algorithmic bias, which threatens to re-write the past and foreclose on possible futures when it comes to the representation of people of color.”