Discover why AI’s hunger for training data could revolutionize content creation economics. A fresh perspective on the intersection of artificial intelligence and creative rights, that looks at one of the unique aspects of the current situation: The machines are now part of your audience.
For the latest in AI news, analysis and tools, subscribe to the newsletter!
Transcript
Storytellers are Developers Part 3: The New Audiences Aren’t All Human
John Gauntt [00:00:06] This is Culture and Code, a podcast about creativity and artificial intelligence. I’m John Gauntt. Culture and Code explores innovation across storytelling, technology and audiences to help professional creators collaborate better with AI and each other. Support this podcast by subscribing to our newsletter. Visit our website, Culture and code.io.
John Gauntt [00:00:32] This is Storytellers Are Developers, a five part series that analyzes how the next generation of AI will be trained on professional, user generated, and synthetic content. Storytellers Are Developers maps a growing symbiosis between the creator economy and the AI economy. This show is Part 3 and focuses on AI models as the next great audience for human creators.
John Gauntt [00:00:55] We often look at technology as the medium or platform for making or distributing content, but it’s a true leap to start including technology itself as part of the audience. But that’s exactly the mentality it’s going to take if content creators are to regain control over their destiny. It’s new territory for innovation with commercial and cultural fault lines on the human side, plus technical and even philosophical challenges for those building AI. So how should professional storytellers navigate this thicket when there’s also this same old script between the media and technology industry that both have legacy business models to protect.
John Gauntt [00:01:34] Well, if we’re going to talk about storytelling and AI let’s treat it like a story. And better yet, let’s treat it like a five act play, because this has been my experience across six technology cycles, including AI. Whether we’re talking about the web or keyword search, social media, mobile or AI, the dramatic arc of how media and technology innovation happens and its impact on individual creators, whether they’re working, freelance or working inside. Well, that hardly varies. The drama kicks off with Act One, which is where you have this technical innovation, which could be the World Wide Web or mobile video. It allows entrepreneurs to access, manipulate or create IP and build huge businesses almost overnight. And when that happens, kind of that fusing of the sexy and the successful, the news cycle will hyperventilate. It will write basically two types of stories. Either this is the most amazing thing that has ever happened to media and is going to change everything. Go to the other pole and it says This is the death of fill in the blank and it’s going to change everything. So if you look at some of the early reviews of Napster, of YouTube, etc., you are seeing that either the birth of and the death of type of stories and that’s Act One. And when that happens, that creates tremendous buzz.
John Gauntt [00:03:00] And in Act two, that’s when the incumbent studios and creative IP holders wake up and lawyer up. You see all these angry screeds about how this new innovation has somehow violating the human spirit. There will be sponsored meetings that are going to spotlight unique, precious aspects of human creativity that must be defended at all costs. And of course, you’ll have your obligatory politician threatened to launch some type of public hearing. And that’s the segue that leads to Act Three when Big Tech and Wall Street declare, hey, way, way, way, way, way, way. It’s all a misunderstanding. We’re going to launch a publisher program. We’re going to find some preferred content partners and really show that this is a partnership. And of course, while that’s happening in parallel, the market research industry, its white paper mills, are churning out all kinds of screeds about how this new paradigm is going to create more jobs than it destroys — after a short period of discomfort. And then the dam breaks when a large, well known incumbent will declare, Hey, they’re going to sail into the uncharted waters to a brighter future with a new technology. And they license direct and and suddenly we have one of those technology and media. Peanut butter and chocolate come together.
John Gauntt [00:04:20] Give that enough time and you get to Act Four when most of these early big media players blow through the original payments from the tech partners. While those initial growth forecasts about this new kind of media that’s going to just kill everything else, they start getting kind of wobbly. You start to witness first generation startup insiders begin to cash out. But also in parallel, the experience has conditioned the audience to permanently change their consumption habits. The suddenly you’re not going to sell a ringtone to a mobile user anymore. The music itself is going to cost $0.99 per track, or you might as well just ditch it all, pay monthly fee and have all you can eat. And as those changes in both consumption and economics ripple through the ecosystem, that’s when the layoffs start or they are announced as painful though necessary, restructurings to position us for future success or to right size for the current competitive environment. That’s also when you see a classic media brand either go out of business or become a shell of its former self.
John Gauntt [00:05:28] We arrive at Act Five, where the creative technology and platforms are no longer innovation but table stakes. Layoffs accelerate. Someone fields a viral hit that’s native to the new paradigm, and everybody, whether they’re media or tech, declares, Hey, we saw this all along. So it’s arguable. Last year, 2024, the curtain rose for Act Three detente between the media industry and technology industry. Because licensing deals between creative IP owners and AI companies hung thick in the air. Google and Meta were banging on any Hollywood studio door that would open to try to license their content for AI training, while OpenAI secured a raft of publishing A-listers to train ChatGPT. You’re talking about News Corp., Axel Springer, Atlantic, Vox, Financial Times. Image platforms followed with Shutterstock, Photobucket and Flickr, and accompanying all of them. Were all these press releases gushing about the possibilities, especially with this new technology. But we know the real driver for Act Three is that IP holders eventually have to make a bet because they’re trying to navigate this multi-level prisoner’s dilemma that almost never breaks their way. And that is. Given that the momentum and the capital is on the side of the tech industry, do you take adequate compensation today or do you take a worse deal tomorrow? Do you fight a legal battle and hope to win and collect? Others are going to dither while their content is still being scraped and monetized by AI companies. And then over time, some of the first voices start hinting that maybe the tech industry doesn’t need to license as much content as they thought they would originally. And over time, both technology and media management start converging on the idea that, hey, given the incredible productivity of these new capabilities, the real costs of creating content actually should be zero. Now, this dynamic is played out time and again since the early 90s. It is a slash and burn agriculture mentality on both the media side and the tech side, a mentality taken to creative output. And all the while the actual creators, the people who are making, are being squeezed to make more for less, faster. And if I sound somewhat cynical, is because I’ve experienced and witnessed it happen again and again. When management gets a shiny new toy and decides they don’t need people anymore while their retained consultants and vendors nod approvingly.
John Gauntt [00:08:04] For the first time in a long time, this time might be different for those who create for a living. Step change improvements in the performance accuracy and cost of large language models are becoming more difficult to achieve, and incremental improvements in LLMs don’t grab headlines but cost serious money. It’s not difficult to imagine. Like slash and burn, the old growth forest of the internet has been logged. The number of open source datasets that you can truly use that have not been used is vanishingly small. At the same time, inside the four walls of organizations, everyone holds their data tight to the vest. The tech players are desperate to try to get around direct licensing because not only is it an administrative hairball, it costs actual money. Now, of course, some are claiming that synthetic data will save the day. You know, AI will generate its own training data and that’s a very sexy idea. But that also butts up against the reality that inbreeding has been a bad idea since, well, life on earth began.
John Gauntt [00:09:11] And what’s more, everyone in the media industry now knows their IP can be used to train AI. Now, the vast majority don’t know how it’s used, but that doesn’t matter. Scarlett Johansson’s case just made it impossible to explain away why content owners shouldn’t be compensated for use of their media to train models. And all the while, the fair use clause of Section 230 of the 1996 Telecommunications Act looks more threadbare by the day. So in that sense, content creation should might just have lucked into its craft beer moment. Like raggedy kids stumbling over a winning lottery ticket, the trick will be to stay disciplined, to demand the full value of the reward, rather than to take some discounted cash up front. And that’s going to require creators to think about their expression and IP in new ways. Now that AI has become just as much of a legitimate audience as 18 to 34 year olds. While on the technology side, 1990s idea that digital automatically equals free is going to be tested like never before. I believe for both sides, the winning move is to create an entirely new game. But that will require both sectors to revisit, reevaluate and remix the mental models or paradigms that made them successful in the first place. And that’s the topic of part four of the Storytellers Are Developer series The Power of Paradigms for Good, Bad and Ugly. Stay tuned.
John Gauntt [00:10:48] You’ve been listening to the Culture & Code podcast, Creativity and Artificial Intelligence. You can find us on all major podcasting platforms and our website culture and code.io. That’s one word. Culture and code.io. To support this podcast, subscribe to our email newsletter for a weekly roundup of analysis, news, plus AI training and growth opportunities for creative professionals. It’s all there at cultureandcode.io Culture and Code is a creative studio and professional education provider that’s part of the Augmented City LLC. I’m John Gauntt inviting you to the next episode of the Culture & Code Podcast. Thanks for listening.
