I'm starting a series on the bubbles that AI is helping create. While everyone is looking at NVIDIA, Anthropic, and other players sucking up huge amounts of investments and even local resources (like electricity and water), we're also in an era of other bubbles that are being exacerbated by AI.
Today I'd like to talk about Creative Inflation - what happens when creating is as easy as clicking a button.
Now, this is not your typical "bubble". I'm not talking 100% about a financial bubble that is about to pop, though in some cases that is also a risk. What I mean is what happens when large corporations start filling all the typical channels - movies, podcasts, TV shows and commercials, social media - with AI slop and what smaller creators can do to fight it.
I first learned of the idea of Creative Inflation from Charlotte Mair, Founder and Managing Director of award-winning culture and communications agency, The Fitting Room, who noted that the decade of the 90's felt far more creative than today's era of endless content. She explains in her video that a budget for a music video like Nirvana's "Smells Like Teen Spirit" was about €50,000, meaning "every creative decision had a financial weight".
Nowadays in music, advertising, and Hollywood, money can feel like it's growing on trees. This is because every moment of content is repurposed, repackaged, and republished on 15 platforms and is analyzed to bits through automated A/B testing, algorithms, and other methods to squeeze every last drop of optimization, which ultimately pad business executives' bonuses. "Creative" work gets a huge budget, while most creatives still seem to be struggling to pay their rent.
AI extends this inflation by allowing anyone, as Charlotte's video shows, to create 50 "variants" in the blink of an eye of a graphic, or in the Superman Moustache-gate example, to save $24,999,500 on fixing an actor's appearance. If you don't know about this story, his moustache was replaced by a fake upper lip using 25 MILLION dollars worth of CGI budget. Then, some guy with a €500 computer and a GenAI tool did a much better job.
Of course that 25 million dollars represents quite a few jobs in CGI that I think are invaluable, and quite frankly, can't be replaced by an AI. That's because an AI doesn't love superheroes the way some nerd who has spent their entire life studying to become a CGI artist and work on superhero movies does. But now that we can do 25 million dollars worth of CGI with 500 dollars worth of technology, where do you think the rest of the movie budget is going - to make the movie better? Or to pay the executives and top-billed talent?
Movie budgets keep getting bigger and bigger, but the distribution of the profits is creating a system where eventually, something will have to give. We have seen in recent times multiple big-budget movies that have flopped at the box office. Of course, some of this is down to creative accounting, but eventually, a house of cards will topple, and I worry that AI may increase the speed at which this happens. AI could do this by writing stories that are less resonant, creating unappealing visuals with an uncanny valley feeling, and even just turning the audience off because they know it's AI.
So, why is this an issue? Because we're getting films (and even TV shows) that are less and less creative even though they cost more and more. 19 of the top 20 most expensive films of all time are sequels. Of course, some breakthrough releases surprise everyone and create a lot of chatter due to offering something completely new (like Sinners), but even before AI, as social media and other technological elements of our society have started to influence art, it has started to feel like everything looks great but has no meaning.
Now, how can AI help fix this?
Smaller filmmakers who have worked hard to be intentional now also have access to tools that can increase the impact of their budget while respecting the team that they work with.
And while I understand plenty of creatives are 100% against AI, these smaller filmmakers are more likely to use it intentionally because there is more riding on their wins and losses. Something creative accounting can't account for.
Intentionality might look like:
This is just an example - but can also be applied to other industries! There's always a way to use AI to make your work better, and even look more expensive on a smaller budget, but not replace the human touch entirely.
In an example outside of Hollywood, but still in the remarkably lucrative field of podcasting, some companies have felt the urge to develop AI-generated podcasts as a business line, to thunderous applause (not).
The Washington Post is privately owned by Jeff Bezos, so their financial statements are not published publicly, but with somewhere between 77 and 100 million dollars lost in 2024, 100 staff members laid off in January 2025, and their lowest subscriber count in 55 years, it's clear that they need to make some changes.
WaPo tried to roll out personalized AI podcasts in late 2024, but within 48 hours, they were struggling with huge amounts of disinformation, which of course, damaged the reputation of the Post. This was critiqued by Bridget Todd and her producer, Mike on the podcast There Are No Girls on the Internet for being "news flavored content without recognizing the special considerations that journalism needs to adhere to, to make sure that readers and listeners trust it." Basically, it's not good.
In my research, I couldn't find the exact investment into AI podcasting, but I have broken down the costs of human-led podcasts vs AI-created ones.
AI podcasting, according to estimates based on other, similar platforms, costs them about €1 per generated episode. They charge anywhere from $4-14 per month for a subscriber. They, presumably, make money on advertising, too - and this is normally priced on a cost per listen basis (or is sold as such, with an estimated return). I think you can tell where I'm going with this.
As far as I can tell, this is meant for a 1 time listen - 1 listener, 1 podcast episode. I'm fairly certain they have more than 400k unique monthly listens to their human-led ones, with Post Reports ranking #2 in the Apple podcast charts, a platform which boasts more than 28.5 million listeners.
This means, the cost per listen of AI podcasts is way higher. So the cost of advertising through AI podcasts will be much higher. And, I'm not sure paying subscribers want to hear advertisements, anyway. So, where's the end goal?
Probably Jeff Bezos wants your data, like when he bought Whole Foods. That's the only guess I can make.
So, not only does this model of AI podcasting make no sense, but in order to make sense, it would cause advertising costs to go up. Technically podcasting's bubble already popped in 2023 and they're in a maturation phase now, but AI could throw a wrench by inflating the number of podcasts out there, again.
WaPo's gamble on AI may even topple a well-respected journalistic entity that has already been struggling. Either way, we may have less access to both GOOD podcasts, and terrible ones!
🛑 Jeff Bezos, please stop! We are so very tired of you.
Now, how can AI help fix this?
Smaller podcasters can use AI to help them automate the tasks that drive up the costs of their production:
Like with film, there may be ways to lighten your load and still produce top-quality human-made content. Basically, use AI for the business part.
Additionally, you're already seeing podcast publishers like iHeart put "Guaranteed Human" at the beginning of their episodes to indicate that you can trust their programming. This is a good way to indicate to viewers that you are a real person, even if AI is helping you to manage your time.
No one loves commercials more than people who have worked in marketing, like me.
The thing about a good commercial is that it can actually have a huge impact on your business, not just in this quarter, but for years to come.
I still remember some of my favorite commercials from childhood - from Yop's "Me Mama" to just so many Concerned Children's Advertisers' public service announcements which played during kids' shows in Canada. Commercials have made a real imprint on my long-term memory.
But commercials aren't what they used to be.
Now, they rely more on elements like celebrity endorsements than pure storytelling, making commercials more and more expensive. Sure, they do generate ad revenue (especially those commercials played at the Super Bowl), but the cost is astronomical for something that isn't really that memorable long-term.
In Europe, McDonald's faced backlash this year for their AI commercial in the Netherlands, and ended up pulling the advert in response. Conversely, Intermarché, a French grocer, received an unprecedented international response to their advert, made with a team of 80 people over a year.
Note: also not without controversy. The story may have been stolen from a small creator, though it's also possible that the story was not so "unique" as to be impossible to recreate entirely independently. For example, it's also not too far away from the Grinch, on the last page.
These are both examples of "big budget" commercials, because they both would have played at expensive times of day, in multiple channels, and come from companies with big budgets. But one of them was created with significantly less intentionality, and was significantly worse.
Nowadays, even small businesses are making "commercials" for YouTube, and salespeople are using video to reach out to prospects.
While AI might seem like a quick ticket to producing personalized content at scale, using it intentionally matters so you don't get lost in the spam folder, or clicked away after 5 seconds.
Some platitudes to keep in mind:
If you check out this recent conversation around email marketing on Reddit, you'll see a similar trend: using AI to help speed up processes that take you time (list segmentation, amongst others) is a much better strategy than giving the reins to AI on the creative part.
I don't want to discourage people from using AI in their creative work completely. I run an AI consultancy, and it benefits me the more people who are interested in learning how to use AI.
However, we need to stop devaluing our audience's experience by creating content that isn't intentional, isn't resonant, and doesn't return a long-term ROI. Lots of people are seeing small successes in throwing AI Agents at content creation right now (with less success than a year ago, I'd argue, since the market is more saturated), but what happens when it's every piece of content we're fed?
What if we never have another Kurt Cobain?
If you're interested in investigating how to use AI to support your intentional creative endeavours, please reach out.
