The Creator's Gate

perspectives | 2026-05-05 | economyforeveryone

When content gets cheap, the hard part isn't making it anymore. The hard part is getting found, trusted, and paid inside systems you don't control.

One small action: Pick one platform that controls your primary discovery path and check its current policy on AI labeling, training-data opt-out, and human review.

AI didn’t create the gate that controls who gets found, trusted, and paid as a creator. But it made that gate a lot more important and a lot harder to get through.

What’s happening

Making content got cheap. Images, text, music, voice - all of it dropped from thousands of dollars and days of skilled labor to cents and seconds. That shift has happened and it is not reversing.

When production gets cheap, the scarce thing is no longer craft. It’s distribution and trust. Whoever controls ranking, discovery, and the labels that let audiences know what they’re looking at - that’s where the power went.

For creators, that means the hard part moved. Making the work was never easy. But making it was something you could control. Getting found, getting paid, and being believed? Those have always depended on platforms and institutions. AI just made it clearer how little leverage most creators have over any of that.

Why it’s happening

Here’s the mechanism, as plainly as I can say it:

Platforms - Spotify, Amazon KDP, Getty, YouTube, Google - controlled the distribution gate before generative AI arrived. Getty and Shutterstock were already the chokepoint for commercial imagery. Spotify already mediated who got heard. AI didn’t build those gates. It added volume, dropped price floors, and created new consent problems on top of a structure that was already concentrated.

Then the consent problem arrived. Decades of creative work - images, books, voice recordings, songs - was pulled into AI training pipelines. Often without asking. Often before any opt-out framework existed. That work is now the foundation of tools that compete directly with the people who made it. Creators didn’t consent to becoming the raw material for their own replacement.

The legal question of whether that constitutes infringement is still unsettled in the US and UK. Courts haven’t ruled. But the sequence is clear: build first, treat consent as a PR problem later.

Creators end up paying twice.

First through economic displacement. The price floor drops. Commissions dry up. A survey of UK illustrators found more than 32% had already lost work to AI, averaging over GBP 9,000 per affected artist. For translators, the risk is higher - one projection puts 56% of revenue at risk in that role by 2028. Entry-level gigs disappear first, because “close enough” gets cheap enough fastest there.

Then through something harder to measure but just as damaging: authenticity collapse. Once people stop trusting what they’re looking at, creators have to spend energy proving their work is human. And the tools being used to flag AI-generated content have documented false-positive rates. A photographer was disqualified from a contest after their actual work was flagged as synthetic.

Both are documented. They’re also different problems with different remedies.

What good looks like

The ask isn’t “stop AI.” It’s: govern the gate that got more important when content got cheap.

That means three things at minimum.

Consent before training. At minimum, creators need an enforceable way to say no - not a voluntary registry that some platforms honor and others don’t. The EU AI Act moves in that direction through a machine-readable opt-out obligation, though enforcement is not yet proven. The US has no equivalent. That gap matters.

Provenance at the point of discovery. AI-generated content should carry a label where audiences actually see it - at the listing, the search result, the stream. Not buried in a description field. Not disclosed only in a press release after backlash. At the point of discovery. This isn’t about punishing AI content. It’s about letting audiences actually choose.

Contestable ranking. Platforms that control a creator’s primary discovery path should have auditable, appealable systems for ranking and takedown decisions. Right now if anything is available at all, it often looks like a form, a delay, and no clear explanation. That’s not meaningful recourse.

Some of this is moving. The WGA negotiated hard floors in 2023: AI can’t be credited as a writer, companies must disclose when materials given to writers were AI-generated. SAG-AFTRA won explicit consent requirements before studios can replicate a performer’s voice or likeness. Those are durable gains inside specific contracts.

The gap is everything outside those contracts. Non-union workers, pre-production AI use, training on existing work, the full creator economy that doesn’t have a guild at all.

What to do

If you’re a creator:

Build at least one channel you own. A newsletter, a direct booking path, a community that doesn’t depend on a platform’s ranking algorithm to survive. When the algorithm shifts - and it will - owned channels are what’s left.

Strengthen signals that are hard to synthesize. Direct relationships, documented process, named collaborators, revision history. These resist substitution better than polished output alone.

When you lose ranking or get a takedown, ask in writing for the specific reason. Document the response, or the non-response. That record matters if accountability catches up.

If you follow this stuff at the policy level:

Push for consumer-facing provenance labels. A disclosure buried in a description field is not a label. The minimum is disclosure at the point of discovery.

Support opt-out with teeth - enforceable, not voluntary. Monitor whether the EU AI Act’s opt-out requirement actually gets enforced. Advocate for equivalent US legislation.

Back contestable ranking. Platforms that control primary discovery should have auditable, appealable systems. That’s a regulatory ask, not something platforms will build voluntarily.

How to talk about it

This is a governance argument, not an anti-technology one.

The line that keeps landing: “I’m not mad that tools exist. I’m mad that my work became raw material for a competitor - without consent, without a deal, and without a label that lets audiences choose.”

That’s the grievance. Not that AI can make images. That the value of decades of creative work got absorbed into systems that now compete with the people who built them - and the people who control those systems captured most of the gain.

Consumers do benefit from cheaper content. That’s documented. The question is who’s being made to finance that benefit: the labs and platforms that captured the economic value, or the creators who didn’t get asked?

When production goes cheap, trust gets expensive. Right now the bill for trust is being sent to creators and audiences instead of to the platforms that control the gate.

In April 2025, WorldCon - the organization that hosts the Hugo Awards - disclosed that ChatGPT had been used to vet over 1,300 panelist applications. The task was framed as information discovery before human review. The community response was immediate. Authors withdrew from the program. Three senior administrators resigned. The chair issued a public apology.

The specific grievance wasn’t that AI made wrong calls. It was that a tool trained on writers’ work had been used to decide which writers got to represent their craft community. The institution used the product of writers’ labor to evaluate the writers. The backlash was emotional. But it pointed at something structural: when institutions can offload judgment to automated systems, the people whose work built those systems have no recourse.

That’s the problem underneath the five different creator grievances - unconsented training, labor devaluation, discovery flooding, authenticity collapse, institutional betrayal. They’re not the same problem. Each has a different mechanism and a different fix. But they share a root: the gate got more powerful when content got cheap, and the people who make content have very little say over how the gate operates.

Governance is what closes that gap. Not reversal. Governance.

One steady action this week

Pick one platform that controls your primary discovery path - your main search feed, your distribution service, your marketplace - and find out what their current policy is on AI content labeling and training data opt-out. Write it down. Then check whether what they say matches what they actually do.

Action ladder

Short term

Creators: Audit one platform you depend on. Check its policies on training-data opt-out, AI labeling, ranking appeals, and takedowns. Keep a written record of what it says and what you observe.
Audiences and supporters: Back one creator through a channel they own - newsletter, direct subscription, direct purchase, direct booking - so your support is not filtered entirely through the platform gate.

Medium term

Creators and creator communities: Coordinate around a small set of shared demands: visible AI labeling at discovery, real opt-out mechanisms, and appeal paths for ranking or takedown decisions. One person complaining is easy to ignore. Organized creators are harder to dismiss.
Institutions, festivals, and marketplaces: Require provenance disclosure and contestable review in the spaces you control. If you host competitions, directories, grants, or discovery systems, make the rules visible and appealable.

Long term

Policymakers and regulators: Push for enforceable consent before training, consumer-facing provenance labels, and contestable ranking systems for platforms that control primary discovery.
Creators, unions, and trade groups: Build durable standards and bargaining power that do not depend on voluntary platform promises. If the gate is structural, the response has to be structural too.

Back to blog