AI’s Test in Non-Fiction Storytelling. How it holds up.

Dr David Dunkley Gyimah
10 min readOct 15, 2023

--

Dr David Dunlkey Gyimah at the British Screen Forum

AI being used for non-fiction sounds like a contradiction in terms, says Jon Gisby, Chair of the influential British Screen Forum. He’s right. I don’t disagree.

AI has come to be associated with non-fiction storytelling, particularly sci-fi movies, such as Ridley Scot’s Blade Runner, Spielberg’s AI and the forthcoming Gareth Edwards’ The Creator. But factual stories?

Titutar Teacher, After AI and Blackruism, Gen AI images usually exemplify Sci-fi genre like these I made

We’re thirty-two floors above the ground in the heart of London’s financial district which yields this spectacular view. Last time I was this high up was at a creative sitting for the Olympic 2012 bid in Canary Wharf. That’s another story. This week it’s an opportunity to share research into AI’s use in creating historical or contemporary stories.

“Flash growth”, that’s what AI expert Professor David Shrier in conversation with World Today Editor Roxanne Escobales refers to AI’s current adoption. As if almost overnight (though yes, it does have a history) and from nowhere it’s opening a constellation of conversations about its application.

Its many uses also subscribe to what we might call non-adjacent creativity. For technologies such as AR/VR — they’re extensions of previous media behaviour, in the same way broadcast news emerged from cinema lifting terms as well such as producer, director and editor.

AI upends that! It represents opportunities we’ve not discovered yet, alongside the perennial problems, and often in the hands of the same actors, that need solving.

It frames conversations that can seemingly be annexations of the world BAI (Before AI) whilst After AI, and by trial and error, entrepreneurs discover new land to settle. Kai Fu Lee’s book AI Super Power exemplifies the latter in his writing about China’s AI explosion. He writes.

China’s market-driven entrepreneurs faced no such dilemma. Unencumbered by lofty mission statements or “core values”, they had no problem following trends in user activity wherever it took those companies.

Among the many then grappling with this new in-your-hand tesseract, my own concepts emerge from personal circumstances. It’s a simple, yet counter intuitive idea which merited prototyping to see its potential.

As with many ideas their eureka moment is the consequence of a lengthy gestation, which if you’re allow me I’d like to expand upon for a few minutes.

The Multimedia Gold Rush

Issue One — 2004.
International award winning Viewmagazine Website, with various web pages in 2005

In 2005 I built a platform Viewmagazine.tv. I’d reached the end of broadcasting (BBC, ABC Channel 4 News) which had served me well and brought accolades from industry.

But it was evident there was a gap in storytelling, diversity and lived experiences I was familiar with. How could I make use of my skillset, from BBC Reportage, Newsnight to freelance reporting from Apartheid South Africa, to capture untold stores that weren’t being featured in legacy media?

My reel circa 2005 of things I’d done in broadcasting edited on my second mac book.

Viewmagazine.tv was a labour of love, assessing user activity and where they would take me. It was photo-driven delivering a genre of story form called cinema journalism which, before the hair on the back of your hands stand up, has nothing to do with sensationalising news. From the magazine video sprung from photos when the user hovered over them.

This all sounds quaint now, but in 2004 Youtube was still yet to greet the world and Internet speeds were running around 2mb — barely enough to watch a video without the interminable stutter.

Months of coding, and designing resulted in the above from an earlier 1999 prototype (below). The back end used compression technology that facilitated ease-of-video-viewing. But that wasn’t the killer app.

Viewmagazine Website in 1999

Inspired by the film Minority Report, I asked what if you were watching a video and you could seamlessly link from one story to another without a glitch.

I read somewhere at an experimental screening of The Matrix, cinema viewers could control the narrative by collective voting, so one theatre would be seeing more aerial Kung Fu, whilst an adjacent theatre would be enveloped with Jungian philosophy. It was possible and I called it video-hyperlinking — a no brainer. Today under AI a film’s narrative can be controlled with a myria of gestures and responses e.g eyes, heartbeat etc.

The Economist, Sunday Times, and Apple called. And then the biggest shock, the US’s respected global innovation in journalism award, the Knight Batten Award, would literally call my mobile one evening whilst friends and I were out eating at Ozer, that exquisite Turkish restaurant off Langham Place (Central London).

“Hello is this David”, the caller, asked with a general US accent. “We were wondering if you were able to to come to Washington next week. You’ll be giving a talk to the media and academia as part of the Batten Award at the National Press Club, about your work. Well fly you over and put you up at nearby hotel”.

Viewmagazine beat legacy media such as Newsweek and CNN and was hailed by the judges as “This interactive magazine foreshadows the future with its use of hip new story forms and highly video-centric Web tools.” Several years later the Guardian would win it proclaiming they were the first British outfit to win the US prize.

From then on, further unencumbered events followed. Being named Asper Visiting Professor for University of British Columbia and spending a deeply enriching experience with faculty and students. Launching a StoryLab in a university and pairing up cohorts with mentors from the Guild of Entrepreneurs which explored AI.

Then relaunching a new AI-inspired programme, and then this time last year being invited to join Google’s small team in Paris poring over the best innovation in Europe who would qualify for €150K. The level of innovation and inspiration, several in AI, sparked several other ideas.

Presenting at SXSW, EU Xchange and Apple’s flasghip Store in London

Twenty years since the Knight Batten Award countless new storytelling platforms have been launched and yet there’s still a large hole, unfinished business in the narratives that remain buried, untold.

As one of 12 national artists in residence at the Southbank Centre, its artistic director Jude Kelly CBE arranged for multiple award winning film maker mark cousins and I to spend some days together exploring each other’s work.

This is what Mark would comment on what he could see from portfolio.

More recently this unfulfilled hole was better eloquently expressed by David Olusoga’s (2020) James MacTaggart Lecture at the Edinburgh TV festival .

This then brings me to the present, and what feels creatively like 2005 all over again, and how my father’s life would serve as a platform for AI-Non Fiction Storytelling. In many ways the story is an allegory for diversity, and an innovative approach to storytelling that merges contemporary history and AI.

Dad’s Story

Dad and friends 1955 in London.

My father (far left) was a charismatic figure. I, unfortunately discovered the full extent of this late. I knew little of his past, which wasn’t helped when we ceased to communicate with one another for fifteen years.

I’d spent my teenage years in Ghana under his care, having had a fairly torrent time with my siblings in the UK hoping from one foster parent to another. In Ghana, after eight years of boarding school I wanted to return to the UK and I did. But I defied his one and only wish to study to be a doctor and return to Ghana.

Dad’s history reads like something out of a novel. He was a policeman in the Gold Coast. In 1955, two years before the Gold Coast would gain independence to become Ghana he and friends travelled on scholarships to the UK.

It was nothing like he’d imagined; the colour bar, unemployment, housing issues were deep challenges. He would abandon the idea of being a policeman in the UK ( which conjures up a vision of a 1950s BBC Luther) and try his hand at several unskilled jobs, whilst studying.

For new Ghanaians seeking to make South London their home, it was not uncommon to be pointed to Edward (dad). He was a networker, regularly held parties; something he continued in Ghana hiring bands like Osibisa, and trying to book Fela Kuti to play in his club. When Ali fought Foreman in Zaire, he popped of to watch the fight live.

His return to Ghana in the mid 70s with us in tow was replete with story arcs. He was being courted as a politician, and it was not uncommon for military to salute him in public with the expression, “Chairman”.

He would survive a military coup in Ghana stripped of his business and assets when entrepreneurs were being targeted. How could I tell his story, when video and photos were scant, but a suitcase of letters he left behind provide empirical research of his life.

Chairman is representational of the many West African and African stories in the diaspora. This year was the 75th anniversary of Windrush — that momentous realisation of the journey of people from the Caribbean to the UK. It’s a huge milestone.

Within that symbolism lies the story of African people like my father, which seeks greater oxygen. Today, in the UK there are more Brits of African (W. African origin), than Caribbean backgrounds.

This is where AI became the solution. It’s deployment has a particular resonance around the absence of media. For instance between the 1930s and 1960s the Laval Decree across Francophone Africa prevented any Africans from filming.

Using Gen AI, I could create images and video based on research whose results were so convincing that my family couldn’t tell the difference and that Dad at 25-years-of-age were AI-generated. Here’s that video in which about 40 percent is AI generated.

The feedback on social media and from industry experts has been nothing short of encouraging, caveated with obvious concerns around whether this approach has ethical concerns which I address. Verification of one’s reality will require standards, much in the same way 19th century impressionist painters reframed their reality within Realism (Art).

Whilst its nigh impossible to know where this goes next, it’s clear this is approach to non-fiction storytelling is in its infancy and advances in prompting (the language for generating images) will only become more advanced in providing more sophisticated assets. In the next five years text-to-image-to-video will become more prevalent.

It’s low bar entry will have a significant impact in filmmaking, and no doubt as always is the trend be adopted in parts or wholesale by professionals.

Collaborating with researchers, we’re also finding ways of bringing inert historical research alive. Next month, a leading international audio and radio journal publishes a special edition of a century of BBC radio featuring research I conducted into BBC radio in London aimed at African and Caribbean people.

It follows an archive find from a show I presented in the 1990s on GLR. Working with a professional archivist Jose a global body would digitise then find classifying it as historically important. Interviews from the archive, such as Eartha Kitt and Fela Kuti are being given the AI-Nonfiction treatment.

However it’s not all good news. Presenting at the Screen Forum too was Ben Keen an Independent Analyst in Technology, Media & Telecoms. His empirical evidence on current investment in TV and film in the UK and investment over the years reveals an existential crisis. The decline in funding TV and films and inward investment from abroad is spiralling downwards.

This could hamper any successes in securing public and private funds for producing films, period, let alone innovation using AI. That said, a considerable amount of innovation in AI filmmaking, particularly in promoting Sci-fi films, with its low to zero funding, is happening outside of mainstream. Its creating something akin to the rise in Instagram photo imagery or TikTok videos.

Thus far I’ve focused on story, but there’s a matrix of AI-influences awaiting us that I believe will, in less than a decade, turn TV and news upside down. I’ve said things like this before based on trend extrapolation and systems thinking.

In 2006 I talked about streaming from home across the network — something Apple picked up and featured on its front website. I’m convinced, and I had an artist visualise this, that within a decade, everything from relationships, seeing fresh directions with stories, and being able to fact check in real time so viewers see a lie as its being made, will be evident.

What Apple showed from my research almost twenty years ago, and the future of TV News with mutiple visual analysis.

Is there a market for factual stories using AI? In as much as it adds value, yes! If you’re interested in getting in touch you can find me at Gyimahd@cardiff.ac.uk

Here for what industry says about my work.

I’m due to speak at venues in the coming week which I’ll advertise on my Linkedin and twitter feed.

For more on the Chairman go here to the website

Thank you to Pete, Craig and Jon from the Screen Forum.

--

--

Dr David Dunkley Gyimah
Dr David Dunkley Gyimah

Written by Dr David Dunkley Gyimah

Creative Technologist & Associate Professor. International Award Winner Cinema journalist. Ex BBC/C4News. Apple profiled Top Writer,

No responses yet