When AI can produce endless content, originality becomes a competitive advantage for content writers. In this post, I explore the research behind this shift, and I’ll show you the tools I use to find ideas, research keywords, and write new content.

Quick disclosure: some of the links below are affiliate links. This means that I may earn a small commission if you decide to click and sign up, but this comes at no extra cost to you.

Yesterday, I opened my inbox, and the email replies were already drafted in my inbox, ready for me to click send. They matched my voice, my tone, and were ready to send. I hadn’t typed a word.

That’s just where my day started. Next, I opened Instagram. My feed is full of AI-generated videos, ads, and altered photos. And the more I use AI tools, the more they learn a little more about how I write, how I plan trips, what I like to cook.

If you make a living creating content, that should worry you. Ask yourself: If AI can write exactly like you, sound like you, plan like you but faster, what makes your writing worth reading over the AI version?

So I went looking at the most recent studies comparing AI and human content. What I found isn’t what you would assume.

How much AI-generated content is being published?

Remember when ChatGPT just launched in November 2022?

It was fun to use, but it didn’t contribute much to the volume of content published online. We used it to write Instagram captions, but quickly gave up due to repetition.

In just one year, AI-generated articles jumped to a 39% share of newly published content. And by 2024, AI content overtook human-written content for the first time, with 51.1% vs. 48.9% share (source: Graphite).

But it didn’t continue growing indefinitely from there.

Which is the most surprising thing.

In fact, Graphite found that AI-generated content has stagnated at around 50% share.

The question is why?

One possible explanation is the fact that Google search started penalising sites that post low-quality AI-generated articles. So the incentive to keep publishing AI-generated content at scale quickly became unattractive.

Read here: How I lost 80% of my Google traffic and what I did to recover it.

But here’s what interests me: If AI can produce content faster than humans, does that content actually perform better over time?

How well does AI-generated content rank on Google?

A 2026 survey of 100 companies, reported by Neil Patel’s team at NP Digital, compared the percentage of AI content vs human content produced and the percent of organic traffic it actually earns.

The results show that human-written content only accounts for 14.5% of what’s published online, yet it brings in 87% of the traffic.

That’s the whole argument for human-written content in one chart.

How well does AI content perform over time?

This is one of the most important stats that changed how I plan new posts.

NP Digital tracked three formats of AI generated content (text, image and video) over a full year against a human-written content baseline.

The baseline is indicated on the chart as a dotted line, held flat at 1.0.

AI-written content starts ahead of human content at 1.3, then in month 2-3 falls by 30%, and then down by 90% by 7-9 months, after which it dies off completely.

Text is the only format that starts out witabove human content rankings, and it’s also the one that falls the hardest.

This is an important finding, telling us not to judge an AI-written post by how it performs in the first week, as that’s likely the highest traffic it will ever get.

Does an AI-generated article traffic compound over time?

One experiment published 744 articles across 68 websites and tracked the average traffic per post for five months. This is the summary chart showing AI-genereated content and human-written content over time:

The gap didn’t shrink over a few months, but it widened.

Human content started at 2.5x in the first month and went up to 5.4x by the fifth month.

What does this all indicate? It suggests that human content compounds over time, while AI content doesn’t follow the same pattern. The variation in traffic is very small over the first five months.

So does originality still matter in SEO content?

Yes, even more than it did a few years ago.

Google’s ranking signals moved in this direction: one analysis of the March 2026 core update found it specifically rebalanced toward EEAT (Experience, Expertise, Authoritativeness and Trustworthiness) and content that shows real firsthand experience gained ground over sites posting generic AI content.

In Semrush’s own study of 42,000 blog posts, only 14% of articles that actually rank were AI-generated, and the #1 spot was 8x more likely to be human-written.

Meanwhile, 72% of SEOs in that same study said they think AI content ranks just as well as human content. This tells you that people perception on AI is that it’s actually doing well – when in fact the proof is still missing. So currently, it’s still a live argument a lot of people are still losing traffic on.

But focusing only on originality isn’t sufficient by itself. It still has to be supported by basic technical SEO and a real strategy.

You can write a brilliant, original post that nobody can find it on Google search, so that doesn’t help your website get traffic either.

Writing new content: what I’m actually doing differently now

The first thing I stopped measuring a new post’s success in the first month. Now I check the analytics back every 3-6 months, because that’s where the real gap between original and AI-generated content starts to show up.

Here’s my actual writing workflow, from start to finish:

1. I start with an SEO audit

I open the Semrush SEO dashboard and check my website’s rankings, organic traffic, backlinks, top keywords, and the basics of search performance.

For me, it’s important to know where my site actually stands before deciding what to publish next.

• Organic Traffic is up +550%
• Organic Keywords up by +6.94%
• Ref. Domains down by 13.15% (This is linked to losing some backlinks, which I need to investigate next)

I’m also looking beyond traditional Google rankings now.

My own SEO and AI Search dashboard in SemRush, not a stock example.

2. I check AI visibility, not just Google stats

I run an AI visibility check to see whether my content is actually getting mentioned in AI tool answers, not just ranking in traditional searches.

This month I saw a rise in AI visibility by almost 10%.

It’s still recovering after the massive traffic drop I had with the Gooogle Core update in May 2026. So, I’m watching both traditional search and AI visibility rather than treating either one as the complete picture.

Semrush shows me the exact pages and where my visibility is coming from, including the traffic from different LLMs like Google AI Overviews, ChatGPT and Perplexity AI.

That matters because a page can rank well in Google without necessarily being cited in an AI answer.

And the good part is that you’re not limited to your own website only, you can check your competitors too. Here’s an example of an AI visibility check:

3. Then I plan the next piece of content

Once I know where I stand, I move to planning new content.

I go to Content → Topic Finder and enter the keyword or topic I want to write about. It gives me a list of related topics and keyword ideas, which I can select based on search volume, relevance and intent.

I can also use the built-in Content Toolkit AI features for ideas and titles when I feel more help is needed. But I don’t let the AI takeover the entire writing process.

Want to start creating original content?

I use Semrush to find topics, select keywords and track both SEO performance and AI visibility, and it’s now become part of my regular content creation workflow.

Conclusion

Overall, AI is not the problem.

Yes, it can help you write 50 posts instead of 5 posts a day.

But if those posts don’t bring you more traffic, visibility and sizzle off after a week, are you even creating actual value?

The data I reviewed in this post suggests that on average human-written articles outperformeds AI-generated ones in one month.

Human experience, originality and authenticity still matters when the goal is long-term content performance.

About the author

Alle Ceambur is the founder of Studio Scribis, a B2B SEO content strategy consultancy for SaaS, AI, and tech brands. Before moving into content marketing, she spent years in corporate business development and team leadership, leading a graphics design team, managing client presentations, and serving as Editor-in-Chief of a company newsletter. She holds an MBA in Global Business and a degree in Accounting.