After more than 35 years of operation, TBI is closing its doors and our website will no longer be updated daily. Thank you for all of your support.
Artificial intelligence ‘bias’ highlighted as tech influences content commissioning & creativity
Artificial Intelligence (AI) should not be seen as a ‘perfect oracle’ when it comes to uses across commissioning and creativity in entertainment, with in-built bias highlighted as a major concern here at the Media & Entertainment Leaders Summit in London.
The nascent tech is being embraced by almost all sectors of the entertainment industry from production groups to streamers such as Netflix, but execs have argued that AI should be treated as a supportive tool to current processes, with particular care around the tech’s in-built assumptions.
Frank Spotnitz, CEO at Big Light Productions, said that while he felt “optimistic” about AI in the long term, its more immediate uses had to be closely controlled.
“In the short term I am really concerned about bias and just bad AI. That means badly written AI, bad assumptions and people putting misplaced faith in AI.
“When it comes to commissioning budgets and creative decisions, if you rely on AI like it is a perfect oracle then that is dangerous.”
Mireality founder Maria Ingold added that managing AI’s bias is a key area to concentrate on for those embracing the tech, while European Broadcasting Union’s head of strategy, Jonathan Broughton, said that the impact of “efficiencies” from AI needed to be explored particularly with a view to the tech taking entry level jobs.
“There is a lot of terror around AI and not just in entertainment,” continued Spotnitz, “but in terms of TV drama and production, the potential is great in terms of improving research, the way we construct a show, how we edit and cover mix.
“There are infinite applications that potentially can help you improve the quality of your work – the conversations I want to have are around how AI can serve people, not the other way round.”
Garrett Goodman, VP of sales at AI dubbing firm Papercup, added that his company employs guardrails to ensure audio content is being produced as accurately as possible.
Having secured consent from actors for their voices, his company then ensures they are “compensated” fairly.
“And when we are doing a dubbing project we also have a human in the loop. So yes, the first pass is done by AI but then there is a professional translator who will fine tune the output.
“That mitigates the risk – there are real ways to put guardrails around them.”