The future of content distribution
Is it too late to fight for ownership of art work in the AI era?
Hello gen AI obsessives — we’ve got a couple of cool things to share this month:
Firstly, we now have a discord channel to engage with our community. Join it here. We’ve been using the channel to discuss news, and share tips and advice about generative AI.
Jeremy and Georgia recorded themselves trying to remake The Simpson’s steamed hams meme with generative AI tools. It was a hilarious and cool learning experience — watch the 5min supercut (scroll down for the full-length video)
We’re running a spring sale on our Generative AI Masterclass right now — use this link to get 10% off on the April cohort (or feel free to share this link with your networks!)
We are in the midst of a radical shift in how we create and consume content
Of course, generative AI is a huge driver of this. The response from artists and creators has been mixed, but largely one of alarm and concern about self-protection. There are lawsuits against Microsoft and OpenAI, and many online publishers have entered unionized negotiations about the use of AI in their work.
Now, I think it’s right for individual creators and organizations within creative industries to be concerned about the ways in which generative AI is changing the content landscape, but I’m concerned that their efforts are going to fail. Not because their motivations are unjustified, or that they’re going up against entities that might be too powerful for them — rather, I’m worried that in 3-5 years the information ecosystem will be so profoundly different that these discussions will be completely irrelevant.
I definitely put myself in the artist and creator bucket. Although it isn’t a big part of my online presence, I make work that lies at the intersection of poetry, installation, and performance, sometimes with a procedural twist. I would by no means consider myself a working artist, although many of my friends are. I’m worried that the battles artists and creators are fighting right now are far too heavily focused on today’s reality of social media newsfeeds, Search, online publications, and media platforms. As important as this seems, I’m concerned that grinding to maintain this status quo could actually hurt creators in the long run.
In 2024, generative AI is feeding into an existing shift in online life: Google feels more like a list of ads than useful content, so people are turning to ChatGPT and Claude for answers. Generating images that encapsulate your aesthetic tastes may now be more alluring (for some) than the infinite scroll of Instagram. Cultural commentators from Katie Notopoulos to Cory Doctorow have written convincingly about how broken the internet feels right now. All of this suggests that something’s gotta give. Even before generative AI, it was obvious that content distribution needed to change. I don’t yet have a concrete proposal for how it should look in the future, or a clear forecast on what I think will happen anyway. What is obvious to me is that we shouldn’t exhaust ourselves trying to keep everything the same.
I can give a clear forecast of what will not happen: the future of content distribution is unlikely to be what we have now, but just with a heavy layer of generated AI content sitting on top. What’s more likely is the emergence of a new paradigm — whether catalyzed by incumbents or new entrants — which will better accommodate the unprecedented ways in which we create and consume content.
Just looking at how things are going now, I can quite easily envision a future where people find content primarily through AI models themselves, and not traditional newsfeeds or streaming platforms. Users might describe their preferences through voice, text, or implicit data, and then be served individualized content tailored to their tastes — and artists excluded from these models will therefore be undiscoverable.
I think the temptation to fight for this kind of exclusion is based on the assumption that the only truly valuable ‘art’ comprises of a set of standard, individual assets: an album; a book of photographs; a novella. What if instead we expanded our understanding of what forms creative works can take, and what it means to be a creative practitioner? The future could see artists fine-tuning models on their work, creating prompt-wrappers, or prompt-tuned vectors which perfectly encapsulate their distinct style. Consider a visionary artist like Lady Gaga: instead of albums, she could give fans access to a model that can generate songs in her style, and in infinite variations. You only wanna hear late 00s Gaga? You got it. Or just songs that sound like they were in A Star is Born? Cool, go for it. Most importantly, she would drop new models with new styles that no one has ever heard before, each unmistakably hers, and each with an infinite number of songs contained within.
The key to this being a positive future revolves around consent and value. In this future, how is Lady Gaga offering the models? Subscription? Where do I find them? How do I hear about them? What happens if someone who isn’t Lady Gaga tries to do this?
Personally, I actually wish that OpenAI and other companies provided a way for me to submit my writing, ideas, and name directly to their models. This way, if anyone ever asks for content in my style, or for information about me, they would receive truly accurate and relevant responses. And, content that is explicitly and reliably generated in the style of Jeremy Kirshbaum would tremendously benefit me by increasing my visibility and reach.
I’m saying all this because I believe it’s important to ask questions about the future of content distribution: what will it look like, and how can we make it work for us? Fighting to essentially remove ourselves from future distribution channels is only going to make things more difficult down the line. There’s something to be said for honing your craft around specific tools and formats — this is a critical part of artistic practice. But I really don’t see any way that this generative AI thing can go back in the box. I don’t think there is an option to stop or slow it down. I think our only way forward now is to figure out how it can work for us, minimize displacement, and use the transition as a leverage point to fix some things that were already broken anyway.
So, rather than pushing against the use of AI models as content channels, it might be more worthwhile thinking about how these shifts can prompt new strategies and business models that work for everyone. Currently, a lot of the thinking around this is still rooted in present-day contexts: some argue that creators should be compensated each time their name is mentioned or their style is emulated by an AI model, similar to how musicians monetize on Spotify. Personally, I wouldn’t support this approach — it’s not likely to be financially viable for creators. Spotify’s streaming model demonstrates that millions of listens are required before artists see any significant income. The platform isn’t even an effective marketing tool either. But, unfortunately, these content distribution channels dominate the space at the moment, and advocating to be locked out of something like Spotify wouldn’t be helpful.
Content distribution as it stands right now is not beneficial to artists. Our energy should not be wasted trying to maintain these systems — we should be thinking about how to build new ones. We need to envision AI-first distribution platforms (or whatever comes after ‘platforms’) that actually value artists’ work; not because we should swallow whole what AI companies are feeding us, but because AI is absolutely going to change the way we value creative work. Not only is it impossible to keep things how they are — we wouldn’t even want it. Pay-per-mention (or similar approaches) will only shove demonstrably ineffective present-day strategies into emerging technologies.
We should put as much effort into defining what the future of content distribution should look like as we do defending creative value in the present, because it's possible that what works for today could actually be counterproductive later down the line. Generative AI does not represent the end of creative expression; it represents an opportunity for carving out new ways of making and sharing creative works.
We don’t know what this positive future of content distribution in the generative AI world looks like, but we do know we want to help build it. If you have ideas about this, please come shout them at us in our Discord.
If someone forwarded this email to you, that’s great — they did the right thing. Now you can do right by all of us and hit subscribe!
We recorded ourselves trying to generate a classic Simpson’s meme
…and I think we did pretty well with the time we had. In under two hours, Georgia explained the cultural significance of Steamed Hams to Jeremy, who had apparently never heard of it (weird), and together they generated some whacky animations WITH a backing track.
Watch us use a range of tools including Pika, DALL-E, and Suno AI to both fail and succeed at what we wanted to generate. We were very much experimenting and improvising — if you have any interest in figuring out how to generate video in certain styles, this might be a good watch for you!
This video is an hour-long deep dive. If you’d rather just a quick hit of content, we also made a five minute supercut of what we made and how we made it. Enjoy!
Enjoyed reading this one Jeremy, unique take and the futurproofing argument makes complete sense. Devansh