Emerging Areas of Complexity and Future Scenarios in Screen

About

Transcript of presentation given to students at the National Film & Television School — March 31, 2025.

100 students present. Presentation recorded ahead as the Foresight Lab was facilitating a Foresight Board meeting the same day.

Invited by Cristen Caine, Jac Sanscartier shares out the core findings of the Introduction to Moments report published December 12, 2024.

Some location identifying information has been removed. It has been adapted and edited for reading.

Transcript

Hi everyone. My name is Jac. I'm a research fellow with CoSTAR Foresight Lab. I’m based at Goldsmiths, University of London. Cristen asked me to go through some of the foresight scenarios we published back in December, as they helped to expose possibilities in some of the dark and challenging, gridlocked issues facing the creative industries.

A bit about my background. I was a foresight strategist for over five years before I joined CoSTAR, and for anyone who doesn’t know foresight, it's a very interdisciplinary field that helps to imagine plausible and preferred futures. Usually foresight strategists are hired in situations of great complexity and uncertainty where there's wicked problems between different systems — what is happening politically, economically and technologically. They especially deal with problems related to technological adoption and how that might affect our lived experience in the future, trying to help people think critically about decisions and what the future consequences might be.

The way we structure this process is through an engagement with leaders in the sector. Our Foresight Board is a collection of leaders from gaming, VFX, machine learning, really across the board [in the creative sector] that represent Britain, because CoSTAR is a UK-based infrastructure project. They reflect creative convergence which is the direction of travel the sector has been going in, sharing in knowledge, skills and not being so siloed. So they helped us think through what our priority areas should be at the start of our foresight study, and when we published the report in December, we wanted to approach this with a perspective of what are the critical areas of inquiry that we should root all our work in at the start of the program. The program is five years long and we're building the foundations of our inquiries.

What you'll notice is these issues aren't entirely futuristic, but they are going to affect the future of the sector quite intensely, and they're very relevant to this moment. And already, while we wrote these scenarios last year, we're already seeing some of them come true quite rapidly.

So I’m going to go through the three emerging areas of complexity that we rooted all our scenario work in.

First is advanced machine learning applications. When we talk about artificial intelligence, we're really talking about advances in the underlying technology, which is machine learning. And this area of complexity deals with the social and economic conflicts [those technologies] pose. You have these challenges at the heart of the industry, where people are feeling unsure around how these tools will change the creative process and how it might change creativity. And with that, anytime that there's this technological adoption underway, there is concern around how it might affect those softer cultural elements, around how we work together, our role at work — technology can wipe out entire roles. And so that's a conflict. But also another conflict is the economic dimension of it. Automation poses, you know, making many jobs redundant. And when we think about some of the tools where the training data is based on copyrighted work, it's a massive conflict with artists and writers who have made the creative sector what it is, and the issues of fair use, which is a conflict that's being discussed in government right now. So that's our first area of complexity that we wanted to draw scenarios around.

Second is platform dynamics. The media landscape has basically transformed over the last 20 years [from] a model that is quite siloed, to it being all accessed via software platforms. Some of those are streaming services, so Netflix, BBC iPlayer, etc, others are more user-generated. So YouTube, this whole class of different media ecosystems today in terms of how they engage people, who the creators are, and also how that affects the production process as well. The production process for a YouTube creator is a bit different from, you know, the ecosystem of the BBC. And that's  interesting, because it shows us that these media ecosystems are really responsive to the initial designs of them. And so we wanted to understand what is the role of platforms in shaping what is produced, and how we interface with cultural production. So that's our second area of complexity.

Third is worker chaos in the creative industries. So this is a reflection of a wider economic issue around employment, but in the creative industries, it really reflects all the conflicts that are bubbling up, also kind of related to the first two areas of complexity around technological change, around systemic change that has occurred, and people not liking the direction of the industries. There being this massive philosophical divide around how things like artificial intelligence should be employed, and you see massive union activity, more unionisation, especially with animation and VFX workers. It’s a reflection of these intense economic, technological and deeply philosophical questions, challenging people who work in production to think, what is it that we want? Do we disagree with these decisions that are happening, among producers, among funders.

So [for] our first area of complexity around advanced machine learning, our provocation for this is ‘who gets to own machine learning?’ Because it's not that machine learning is bad. It's been present in, you know, any kind of software workflow for a long time, but the problem with it right now is how it's designed, and how it relies on training data.

So the way that we map these scenarios is by mapping these two tensions.

One tension being an uncertainty around how and to what extent will it be applied. Are people going to engage in complex training of the tools? Is there going to be sort of a mass training of people where, sort of like with software development well over a decade ago, people know how to use these tools, and they're [developing] them to create new media experiences. And to what extent are people going to just be using them? Like, to what extent is it just going to be productised and sort of a business as usual situation where they just become the new tools that we use.

Second is ownership. So the ownership conflict is [asking] to what extent is there going to be a grassroots movement that is going to challenge how these technologies are developed, bring in new kind of thinking, new philosophy. When you look at the health of the open source community, that's all grassroots and enabling grassroots behavior. [On the other side] is corporate ownership. So to what extent are corporations going to be the owners of the technology at the end of the day, and the wealth will stick with them as well, and alongside that, gatekeeping. What you see is corporations will say that they are making their model open source, but it's not truly open source. It's restricted in many ways.

So I'm going to go through the [machine learning] scenarios, first is IP fracking. So IP fracking refers to organisations engaging in training the models and owning it in house. When there was an announcement last year around Runway and the Lionsgate partnering, that’s a great example of what will likely be IP fracking in the future. IP fracking is a total commoditisation of media. Organisations use their intellectual property and their data to create these new forms of media, and [the media itself] basically becomes a tool. So all the creative works they've used in the past becomes a tool to push the technology and see where it goes. There's a lot of experimenting happening right now around that. And so this scenario just tracks what that might look like in the future, where the focus goes on managing these complex machine learning technologies and away from our traditional production workflows. Also, that more companies will become media companies, because they have tools to monetise their data, monetise their intellectual property, and create media ecosystems from them.

[Next scenario] is another data training scenario, but more grassroots, so happening on the ground. Data alchemy refers to a punk movement that leads to new forms of training data, and also, new methods in co-ownership. The data alchemy scenario I quite like, because it touches on how everyday artists and creators who are experimenting with this technology might think differently about how to use it. And so this scenario looks at new ways of experiencing media online, new genres being created, which is incredibly exciting from a film and TV perspective. If you look at AI horrorcore, it shows how machine learning technologies perceive people and media and our culture in these really strange ways. And it's subverting this big AI tech of perspective of ‘oh, it's going to be a model for all’ to maybe an ecosystem where you have all these small little tools that have been created by people that are ethically trained and serve the purpose of generating new creative possibilities. So we love data alchemy.

Over to collective intelligence.¹ This refers to special interest groups providing these tools. It's a public initiative, and basically when you use these tools, you recognise that you're tapping into the collective intelligence of the creative sector. They're shared assets. When you use them you're tapping into the gaming ecosystem, the film ecosystem, reflecting the co-ownership of of the training data and those tools. So, that's quite interesting, because it can create these after effects of, okay, when I use this tool, I'm tapping into a collective intelligence. What are the cultural effects of that?

And then over to auto pipeline, this is a productised environment where just more film and TV specific tools, for example, are used to totally change the pipeline. And it just becomes a business as usual scenario, just different softwares.

Moving over to the next provocation, what platforms will facilitate culture. So it's good to think of our current platforms as a type of media ecology. Consider]Netflix, the way it's designed, because it's such a great exemplar of how it's changed film and TV. It's just one example of how [platforms] affect people watching film. You know, when we think about Netflix, we think of Netflix and chill, but there's all these other platforms that have these different ecologies based on how [they’re] designed. YouTube  incentivises really short form watching that you can watch on your phone. TikTok incentivises more of a creator ecosystem where the barrier to entry is quite low and there’s a sense of incentive for everyone to be a creator, and that's kind of how the tool is designed to make it very, in my opinion, a very creator-led system where it's quite easy to get started and share whatever media you're producing.

So we wanted to look at two tensions here. We want to look at design, and we want to look at transaction.

So the design of the platform. To what extent is it distributed or centralised? Centralised being about optimising for a certain metric, like for shareholders, really. So attention would be an obvious one. When you go on many Meta platforms, they're optimising for attention because they're an ad business. Distributed being the focus is to optimize user agency. A lot of what you see in web3 is pretty distributed.

We also wanted to map implicit versus explicit transaction. So explicit being there's some very explicit transaction that you're making to enter the platform. Fees being an obvious example. The other side of the spectrum is implicit. So you are the product, or you're engaging on the platform in a way that might be organic, and that is the exchange that you're making. So on one side of the spectrum, with Facebook, you're selling your data by being there. But also you see this in communities on Discord, where the exchange you're making on those communities is your contributions, specifically in communities that are free.

So going over to the scenarios, creative independent union refers to platforms that create a contractual agreement between clusters of creators, allowing them to interface intimately with an audience that is assigned to them. And the idea is it creates a space outside of the Meta platforms, outside of dominant ecosystems of film and TV, and allows people to generate an entirely different ecosystem, where the creators are more in control, and how they come together is quite unique. It's formalised and you have the audiences paying for exchanges, because you see a lot of signals of people wanting more of an organic relationship with artists, because everything feels quite impersonal now. In this sense, a creative independent union can be the basis for creators taking back control and creating their own ecosystem that might be more equitable. It might be more interesting. It might allow for pieces of work that are reflective of emerging subcultures. With a lot of platforms, the feeling of content is very flat, very globalised. So that's creative independent union.

Over to production salons. So new subcultures emerge online, nurtured and studied in closed spaces. I was inspired by a lot of what was happening in Discord when we wrote the scenario. So it’s a distributed scenario, very peer to peer, but the transaction’s implicit. This is an invite only situation where people incubate these production projects through these closed spaces, online. And so you might go on in these very informal ways. You might be a script writer. Perhaps you're using new machine learning tools or experimenting with different forms of production in this distributed way to create a project. And when you create that project, the group decides if they want to sell it onwards or what they want to do with it. And it's just a push from a lot of the creative behavior that's happening on Discord to a more of a dominant [norm with] people creating high fidelity production studios in these closed spaces. So that's production salons.

[Then there’s] mutual binging, a very dystopian scenario. All your biological markers are used for the sake of personalisation. Some people might find that interesting, like in certain healthcare or wellness contexts. But basically personalisation expands to using all available information about you as a person to personalise media, and machine learning [techniques] would play a role in that as well. And [the platform] would be managed by some kind of centralised control, a corporation that is optimising for attention. For example, it might be selling your data onwards.

Finally over to cosy media clubs. This is a gatekept membership collective run by a centralised body, and is managed by critics. So, in a landscape where media is ubiquitous, it becomes about access, and this would be exclusive access to media subcultures. So  you can see something like this happening in skater culture, where you have to pay to just access this media ecosystem of secret documentaries and interviews that is only available to you because you've been even given access, with the critic managing the feel and curation of that space. So that's cosy media clubs, and it's really just an iteration of a lot of what's going on with tokenised collectives.

And then [the final provocation], how will creative work evolve?

So ethos versus governance, these are the tensions that we wanted to map in terms of production work and how workflows might change, but also how we think and feel about where creativity lives, and how that might change.

So ethos reflects a Luddite versus techno-optimist tension that we want to map. Luddite refers to a movement in the late 18^th century if I remember correctly, of silk weavers that wanted to fight back against the automation of their work. Because they felt it was going to upend their lives — you know, take their jobs — but also lead to products and services that were going to be of reduced quality. The head of public value research [at the BBC], Bill Thompson, clarified the second point to me recently. And techno optimists, when you think of American big tech in general and very futuristic, or just futurisms of media technology, a lot of VR  fits within this, so a [style] of going very extreme, asking where can technology take us, where can that level of experimentation take us, versus a desire to maintain control over a workflow with inherent value in film and TV.

So on the governance angle we wanted to map — what does an intervention scenario look like versus a laissez faire, let the I the market decide situation. And the reason why the government would intervene is because of unrest. So creatives don't like artificial intelligence being deployed. You know, the government AI IP consultation is something that could lead to an intervention. But it could also be on the heritage front, to preserve creative identities that might be eroded if they just let the market deploy any form of technology to replace certain types of work.

So, going over scenarios, first is new vintage. This is a program that preserves a very classic form of the creative industry. So with film and TV, there's all this romanticism. You know, in the 80s and the 90s, the certain media technologies that were used, the approaches, the mindsets, the way of working, new vintage would seek to codify all those elements in a specific group of people in the sector that work convergently across gaming, film, and so on. And it creates a point of differentiation for Britain that they're these experts in new vintage. And you can see this being a valid route because of the amount of nostalgia and a desire to recreate what film and TV used to be. There's really interesting signals around a longing for that, that ecosystem that used to exist. That’s new vintage — government intervention, Luddite scenario.

Over to offline revolution. This is a Luddite scenario. So, what if there's no intervention? You let the markets behave out they behave. You know, there could be an offline revolution. There could be recognition that the further digitisation, automation of our media ecosystems makes people revolt and want to return to a material world. So this means maybe a return to cinemas, a return to old production workflows. And a sense that the digitisation movement created inequity by transferring a lot of wealth to technology owners. A lot of those technology owners living overseas. And there's a sense that if you're using a software, it is a medium of perpetuating inequity or the dilution of the industries. And there's all these poetics that emerge from this revolution. So that's offline revolution.

Next is glossy worlds only. Laissez faire situation, with techno optimism [mindset]. Very quickly, VR, AI on steroids. The real-synthetic binary cease to matter. When people talk about media generated from AI versus a traditional production workflow, camera technology. Well, people don't care anymore. They don't care about IP being from provenance versus a derivative IP that various strangers on the internet have generated. It's all viewed equally.

And then finally, over to hyper-polytechnic. This is an intervention where Britain becomes the factory of the world for all forms of creative media. This is already pretty much the case, but this is just taking it to more of an extreme. All arts and media universities become technical residencies, and the lean is very, very, very technical. I use the term STEAM, which a lot of people don't like, but it's STEM but arts, and that's the vibe when you work in these production studios. I mean, it's already very industrialised, but it becomes even more optimised. Think assembly line, a feeling of mechanical workflow. And for anyone who was left behind by this movement, they might move towards a universal creative income to practice their craft. So this might happen to puppeteers and people where their work might be simulated by new technologies. And they're like, well, this is my life. This is what I do. This my craft, and so I want to continue to do this. It could happen with dancers. It could happen with anyone where the physicality of what they do could be just codified, simulated, recreated, digitally. So that's hyper-polytechnic.

There [are also] visual depictions of the scenarios [for] data alchemy, which I went through on the first provocation on machine learning, creative independent union for platforms, and new vintage [in creative work]. We worked with a speculative artist named Jiarong Yu. She's a doctoral researcher in human and AI collaboration.

Finally, I’ll [briefly review] some discourse with our Foresight Board and what they felt, when we went through the provocations with them.

Around advanced machine learning, being clear about the importance of open source behavior, of clear regulatory guidance, focusing on the emerging people coming from within film and TV that are going to generate ethical and artistic led applications. And to really keep an eye on equity and the equity issues that are going on.

Platforms, mainly just recognising the role that platforms play in the media ecosystem. Creating  innovation programs to catalyse new platforms that are UK based, thinking differently about intellectual property in these spaces, because the way that intellectual property evolves — like viral behavior, for example — is so chaotic and not really well understood. And how value is captured is not well understood.

And then finally, around creativity, understanding, what are the creative jobs of the future? Is film and TV work always going to be considered creative, you know, preserving humane activity in the sector, ensuring that creativity does continue to have a place, and that it doesn't become mechanical, too technical, to the extent that we lose the magic and the value of that we love around cinema. Also, to be aware of transhumanist ideologies and where they might take us.

That's all from me. I hope you enjoy the future scenarios. They are actually fully written narrative depictions and have a bunch of detail in them, so I would recommend you read the report, which I forwarded to Cristen so he’ll send it over to you guys. My name is Jac, please reach out if anything I’ve said sparks some interest. Enjoy the rest of your presentation with Cristen.

1 Collective intelligence is a way of thinking about machine learning, popularly introduced by Holly Herndon.

Creative Complexity Triptych - Jiarong Yu