Moving from pre-production to the production process, the filmmaking process begins – we are presuming in this scenario that a production has been greenlit and the producer has received financial backing to proceed with a project. The length of the shoot, actors’ times and the number of locations can all have an impact on final budgets. Once the product is made, the producer will also negotiate contracts for the grant of rights with various rights owners and provide a package of primary transmission rights to the broadcaster. Contractual negotiations will also be in progress with actors, during which, producer’s use of AI may be discussed.

There may be advantages of utilising AI tools during the production process to help reduce the cost of creating background footage, camera usage, auto framing, and sequencing. This, in turn, may reduce the number of crew required to shoot a film (and associated time and costs).

Those areas that AI could help with include:

– Controlling cameras/lights including multi camera use. AI can adjust lighting and set up multi cameras with the use of just one camera person in a shorter time period. (This uses assistive AI)

– AI-led auto framing and sequencing can shorten the time it takes to crop and sequence shots from feeds. (This uses assistive AI)

– Video creation/Image creation including using synthetic media.

– Editing – AI can be used to automatically select the best footage based on lighting and composition.

– Transcriptions/Translations.

– Paper edit transferred to mock edit.

Risks

– Automation risks mistakes/bias and unexpected results, and a lack of creativity and judgement in shots. Vital that AI is seen as a tool to assist production rather than as a replacement.

– Ethical considerations, especially regarding the impact of efficiencies on the workforce and/or misuse of AI to spread misinformation/ and use of personal and proprietary/confidential data.

Opportunities

– Repetitive time consuming tasks may be eradicated, helping to make the production process faster, more cost effective and accessible to a broader range of people.

– More time allowed for focusing on other parts of the creative process.

– Repurposing of existing archive where anything audio based can be turned into authentic media.

Questions to consider

– Are actors happy with AI technology being used to detect faces to quickly identify scenes and characters? Do you know if the specific technology used retains facial recognition data? If so, are actors happy with this?

– If image creation is used to create background images, how has the content been trained? Is it using licensed images?

– Within the Terms and Conditions of the AI tool you are using, do you know the status of the producer’s copyright once a piece of content has been produced? On initial analysis of the tools that Pact has come across, some open access tools state that they may use producer inputs and outputs to train their AI models and that a producer will be granting the developer a licence to use any inputs and outputs. Are you happy with this and does it make commercial sense?

– Given this granting of a licence, are you in a position to grant the owner of the AI tool a licence to use and exploit the content you are uploading? Consider in particular any third party content.

– Are you happy to lose some creative control if you choose to use AI tools in the production process? Have you looked into the terms of copyright ownership and associated risk of your inputs/outputs being trained without your knowledge.

– What are the broadcaster’s editorial standards on particular content? Would this cover how you treat certain AI tools/techniques used?

Case Study

Ethical decisions Documentary Gerry Anderson: A Life Uncharted – Format Factory
During the production of this documentary about Gerry Anderson, the creator of the children television series Thunderbirds, key ethical decisions had to be taken when using AI processes within the film. Producers were given access to 36 hours of audio archive where Gerry talked intimately about his life story. Unfortunately, the audio quality of those recordings was very poor, so they looked into the possibility of recreating the exact interview using voice synthesis. However, during the evaluation of the synthetic voice tool, made by Respeecher, the producers felt that even if they recreated Gerry's interview word for word, it could very easily be deemed as inauthentic by audiences who didn't fully understand the process, which could in turn undermine the truthful nature of the documentary. Effectively highlighting that just because tech exists it doesn't mean it should always be used.

Return to AI Guidance & Resources