IBC
article here
Alice Taylor, the head of BBC Studios new AI Creative Lab describes a “strong, cautious enthusiasm” among production teams for the initiative.
Alice Taylor has a straightforward view of what she does not want the BBC Studios AI Creative Lab to produce: more AI “slop”.
“Certainly from the creative community, I think fully
generated, not very supervised is what we would call slop. And nobody’s into
that game,” says Taylor, who joined BBC Studios at the end of last year to lead
its newly established AI Creative Lab.
For Taylor, the distinction is between AI as a replacement
for creative decision-making and AI as a set of tools that can support it. “It
is technology that serves the creative process and serves the workflow,” she
says. “Everything that we put out, if and where an element of AI is in that
workflow, it’s going to be a portion of the pixels on the screen. Which is a
very different world from slop that’s out on the internet.”
That approach is central to the lab, which was established
to explore how rapidly developing AI technologies can be applied across BBC
Studios’ global production operation. Taylor reports to Zai Bennett, CEO and
Chief Creative Officer of BBC Studios Productions, and is responsible for
developing the lab’s strategy, building its specialist team and working with
production genres including Scripted, Unscripted, Audio, Kids & Family and
Global Productions.
The lab is not intended to be a separate production unit
that simply makes AI-generated programmes. Instead, Taylor describes its remit
as working with creatives to understand AI and determine where it can genuinely
help them.
“Creatives come first. They describe what they’re
imagining,” she says. AI can help shorten the distance between that initial
idea and a visual storytelling element, whether a pre-visualisation or a shot.
“That can be really sped up, which means they can do more. They can, as we say,
throw more spaghetti at the wall, try more things.”
Taylor joined BBC Studios after almost a decade at The Walt
Disney Studios, where she was Director of Technology Innovation and Head of
StudioLAB. There, she led a cross-functional group focused on storytelling
technology, products and experiences, working with technology companies
including Microsoft, Accenture, HPE and T-Mobile.
Her projects included ‘Remembering’, an interactive
experience on Disney+, work bringing Walt Disney “back to life” for the Disney
100 Exhibition, and an immersive Secret Cinema production based around Guardians
of the Galaxy Vol. 2. Before Disney, she founded MakieLab, a company that
combined digital avatars, 3D printing and traditional toy manufacturing to
create customised dolls and action figures. MakieLab was acquired in 2016.
The combination of creative production and emerging
technology is therefore familiar territory. At BBC Studios, however, the scale
of the brief is different. AI is potentially relevant to almost every stage of
production.
“At this point AI is such a broad term,” Taylor says. “There
are tools that are AI-based that have already emerged for every single step of
the production pipeline, whether it’s whiteboarding through to VFX, post and
marketing, through to further down the content pipeline.”
The challenge for the lab is therefore not just identifying
new technology. It is working out where it is useful, where it is mature enough
to deploy and whether it improves the creative result.
“I’m looking at the existing workflow at every step of the
way, because we have to preserve the creative intent,” Taylor says. “We have to
not only preserve the creative intent, but maximise the creative intent.”
The lab is benchmarking tools as they emerge and assessing
existing production software as AI capabilities are increasingly built into it.
But Taylor says the process starts with the creator rather than the technology.
That principle also shapes the lab’s approach to
AI-generated writing, directing and final-pixel imagery. Taylor is unequivocal
that the human creative remains central.
“One hundred per cent. We are people first, creativity
first, trust first,” she says. “We’re coming at this 100 per cent from our
creative-led opportunity space.”
The lab’s small team is being built around that philosophy.
Since January, Taylor has hired two people; Rish Dhir who’s a Creative
Technologist with experience at Metaverse VR and Red Bull and Nathan Hubble, a
filmmaker who founded automation studio Atta Bouy who’s the Program &
Automation Manager. The ideal candidate, she says, combines technical expertise
with an understanding of how television and content are actually made.
“I was looking for kind of creator-inventors,” she says.
“Folks who have a film and television or creative background in some way, but
also have technology expertise and hands-on expertise with AI.”
VFX, animation and film or television production are
particularly relevant backgrounds, alongside an instinct for modernising
established workflows.
The lab is also intended to operate as a bridge between
production teams and the wider technology ecosystem. Taylor says BBC Studios is
talking to UK-based research and development organisations, although she is not
yet naming partners.
Her previous career gives her a wide network to draw on.
During her time at Disney, she was responsible for relationships with
technology companies and led StudioLAB’s innovation portfolio. She also spent
five years running MakieLab between London and Los Angeles, developing a
business around the emerging intersection of digital content, games,
customisation and physical products.
She continues to track developments internationally,
including in the US, Europe and China. Asked whether her relationships with
major AI companies give her access to technologies before they are publicly
released, Taylor declines to elaborate.
“I’ve signed too many NDAs to answer that question,” she
says. “I think I would say we have great relationships and can stay on top of
the absolute cutting edge of what’s going on.”
For BBC Studios’ production teams, the response to AI so far
has been what Taylor describes as “strong, cautious enthusiasm”. The attraction
is partly practical. Production companies are under pressure to deliver
high-quality programmes within constrained budgets and schedules, while
creators are looking for more time to develop ideas, research and storytelling.
“Our brilliant folks are always asking the question: how can
I tell better stories with the budget that I have, in the time that I have?”
Taylor says.
That enthusiasm is balanced by concern about audience trust.
Taylor says creators want to experiment, but “they’re also hyper responsible”.
The objective is to use the technology “correctly, carefully, without
misleading the audience, without losing trust”.
That distinction becomes particularly important when
considering the audience’s response to AI-generated material. Taylor
acknowledges that there is a spectrum of opinion, including viewers who enjoy
fully AI-generated video. But she sees a significant difference between that
material and AI used as part of a professionally supervised production process.
The latter, she argues, is not entirely new. Machine
learning and other forms of AI have already been used in areas such as
animation and VFX for years. What is changing is the breadth and accessibility
of the technology.
For now, the lab is focused less on predicting where AI will
end up than on testing what it can do for creators today.
“I don’t think AI will create new formats, but I think AI
enables the creatives to have more ideas faster,” she says. “It will throw up
ideas, if you ask it to, but it’ll be humans that create formats.”
Historically, she argues, creative ambition has been
constrained by time and money. If technology can give creators greater freedom
within those constraints, it may ultimately lead to new ideas and new forms of
storytelling.
The first results from the lab’s work will take time to
appear because television production operates on long cycles. Some BBC Studios
programmes already contain AI elements, Taylor says, and work that began after
she joined is expected to start appearing on screen within roughly a year,
depending on the production.
For Taylor, that is the point of the lab: not to make AI the
story, but to establish where it can make the creative process better.
“More is not necessarily better,” she says. “But better is better and we’ve got to focus on that."
No comments:
Post a Comment