2025 Predictions: AI Content Fails

Look. It's Q4. I'm busy. You're busy. Let's cut the longwinded opinion round-ups and focus on what we really want: my reasonably accurate annual predictions of what the next year will bring in the creator economy. For the next four Mondays, I will present my top four 2025 predictions along with the signals I'm seeing that are pushing me in that direction.

In partnership with

I admit it: I feel like at least half of my past annual predictions have been pretty easy pulls.

I have a feeling this is the same for a lot of industries.

Trends are generally momentum-based. If, for example, the numbers are increasing for influencer marketing spend or new features are launching left-and-right for podcasts, it’s easy to say, “Influencer marketing spend will go way up” and “Podcasts will see explosive growth.”

So I’m not going to do that.

This year, I’m going to focus on four predictions that I would consider to be ‘big swings’, then explain myself.

Another admission: I’m also doing this because I need to pre-schedule my newsletters through December due to a heavy workload and holiday travel.

Let’s dive in.

2025 Creator Economy Prediction #1

2024, Dall-E 3

AI Creators Fail on Major Platforms

I can’t go a day without a digital exec telling me end-to-end AI will be the norm on all major video platforms.

The ‘Three-Person Marvel Movie.’ The AI-generated South Park episode. The countless faceless channels churning out true crime shorts.

All signs should point to AI content becoming the norm.

Text-to-video platforms like InVideo are getting really good. Suno can make near-studio-quality music for free. At countless text-to-image generators are starting to not only create hyper-realistic imagery but, more importantly, include the correct number of fingers.

So why am I giving this trend a vote of no faith?

Because the platforms still control distribution.

And, more importantly, you need someone to actually watch this stuff.

If I were at a major platform (as I have in the past) I would be thinking five steps ahead:

  1. AI video creation will be so cheap and easy that anyone can do it. In 2025, if someone wants to launch a channel that looks and sounds like CocoMelon, they will likely be able to do it for a small monthly fee.

  2. When anyone can do it, a lot of people will. Should go without saying, but there are a lot of people interested in making internet cash. If you can spin up a channel and make 500 videos per month automatically for $100, tens of thousands of people will. They’ll do the very simple math: if these 500 videos can cumulatively get a few hundred views on average each, you double your investment.

  3. Creation will approach the level, or perhaps outpace, consumption. When more videos are being made faster than people are watching, which will absolutely happen when content is fully generated by AI if the incentives of doing so are appropriate, many videos will get zero views. This could flood major platforms with slop nobody is watching, which costs the platforms money.

  4. Major platforms will continue refining their AI identification systems and devaluing AI slop in their algorithm. As a way to encourage the continued creation of unique and interesting content that brands love to buy, platforms will have to heavily devalue AI content so the incentives of uploading tons of garbage don’t make sense. The more slop is uploaded, the more they’ll tank the content. I bet some will prioritize features where users can ‘shut off’ surfacing AI content.

  5. Smart creators will figure out where they can incorporate AI to avoid this algorithmic drubbing and they’ll be the real winners. The platforms, believe it or not, make nearly all decisions based on what their advertisers want. Advertisers will want high-end, original, aspirational, and interesting content that aligns with their brand image. This will continue to be driven by human beings. There will always be the IronMouse and Lil’ Miquela’s outliers, but to scale those unique cases requires an incredible amount of thought that, frankly, may even be more inefficient than a human creator.

VCs are clamoring for end-to-end video AI because it, in theory, prints money. Or prints IP. Or prints attention.

But you don’t hear audiences clamoring for it.

End-to-end video AI will continue to be sufficient for very specific genres: pornography, faceless explainer videos over B-roll, rehashing human-made work, and brain rot.

Platforms accept this content, but don’t love it.

And the fact that all of them already have a ‘this is made by AI’ checkbox when uploading video is a sign of what’s to come.

Content created with too much AI will, in 2025 at least, continue to be second-tier content.

Creators who are deftly using AI to speed up workflows and increase content quality, on the other hand, will thrive. And that’s where attention should be paid.

40-55 In US? Add 20% To $200k+ Retirement Savings Today

  • Ensure your savings last throughout retirement

  • Shield your savings from unpredictable market swings

  • Add 20% growth immediately to your retirement savings

Thank you for reading and stay tuned through December for more 2025 predictions! If you enjoyed this edition, give it a share and if you get someone to sign up, I’ll send you my ‘10 Rep-Friendly Ways to Monetize Today!’ deck!

Until next time, protect yo rep.