NOTE:
This article was 100% written by ChatGPT 4.o. It is an experiment, related to this article: “Can you tell if an article/video was AI-created?“
Could you really tell it was written by a robot?
The prompt for this was:
“Can you write a full article about the following topic:
AI-generated articles and videos/actors are getting more and more convincing, human-like. Discussing the current state of the newly created “content” – and how prevalent it’s become in 2025 (majority of newly written articles are AI-written).”
1. Introduction: The content revolution we didn’t ask for
As of 2025, the vast majority of newly published articles, blog posts, and even videos are not written or presented by humans. They are generated by artificial intelligence – text written by language models, speech created by voice synthesis, and faces animated by deepfake or avatar engines.
This shift didn’t arrive overnight, but it feels that way. One moment, AI-generated content was obvious, robotic, even humorous in its awkwardness. Now, most readers and viewers can’t tell the difference – and in most cases, they aren’t even aware there’s a difference to spot.
2. From novelty to dominance: A timeline of acceleration
Between 2022 and 2023, tools like ChatGPT, Claude, and others gained public traction. At first, they were used to assist writers, marketers, and creators – a draft here, a product description there. But as 2024 rolled in, two factors caused an exponential shift:
- Economic incentives: AI can churn out hundreds of articles or videos per day at near-zero marginal cost. No salaries. No schedules. No fatigue.
- Technical refinement: The models became significantly more coherent, expressive, and stylistically flexible – mimicking tone, humour, and even personal quirks.
By mid-2024, media outlets, SEO agencies, and content farms had fully embraced automation. And by 2025, it’s safe to say:
More than 90% of new written content online is generated by AI.
For commercial websites, that number is likely above 99%.
3. It’s not just text: Videos are synthetic, too
AI-generated content is no longer confined to words on a screen.
- AI avatars now read the news, host explainer videos, and sell products — complete with eye movement, lip syncing, and facial micro-expressions.
- Voice cloning creates synthetic narrators indistinguishable from real humans — sometimes even impersonating celebrities, influencers, or past hosts.
- Scriptwriting AI now builds entire YouTube channels: from idea generation and SEO-optimised titles, to thumbnail design and full narration.
In other words: some of the people you’re watching aren’t people. Some never existed at all.
4. The illusion of authenticity
What makes this especially potent is that AI content now feels human. It imitates conversational tone, adds storytelling beats, and even inserts deliberate “flaws” like filler words or hesitant phrasing — to simulate authenticity.
In 2023, AI articles still felt flat or overstructured. In 2025, they “ramble” just enough to seem natural. They “make typos” or “go off-topic” just slightly — like real people. But there is no person. There never was.
And that raises a hard question: if it feels human, and it sounds human, but it was written in milliseconds by a model… does the difference matter?
5. Who benefits – and who loses
AI content makes sense for corporations. It reduces costs and boosts output. Large sites can publish thousands of pieces weekly, saturating every niche with algorithmically tailored material.
Meanwhile, human writers and creators are being drowned out – not just in numbers, but in visibility. Search engines and platforms increasingly surface the most optimised, prolific publishers. That’s almost always the AI-assisted or AI-driven ones.
Smaller independent blogs, niche YouTubers, and genuine experts find themselves invisible. Their quality work simply doesn’t scale at the speed AI does. And increasingly, audiences don’t even notice it’s missing.
6. Is there still a place for human-made content?
Yes – but it’s shrinking, and it’s not where most people look.
Genuine, human-made content still exists:
- In forums and newsletters.
- On personal websites.
- In podcasts with unscripted, wandering conversations.
- In comment sections – when not flooded by spammy bots.
It’s also still present in video essays, reviews, and journalism… but often buried beneath layers of AI-enhanced or fully AI-made clones. You can still find the real thing, but you have to search for it — and most people won’t bother.
7. Final thoughts: The uncanny age
We are now living in an uncanny phase of content history. Most of what we read and watch is synthetic, yet presented as human. There’s no required labelling. No warnings. No disclosures.
In a way, it’s not that AI is “fooling” us. It’s that we’ve stopped caring whether what we consume was made by people.
That indifference may prove more significant than any technical breakthrough. Because once the audience stops asking “who made this?” – there’s little reason for platforms or publishers to return to human creation.
And when everything feels authentic, nothing really is.
Last updated:
Originally published:
