Do AI-written articles get less search traffic than human ones, or is 'less traffic' the wrong thing to measure?

This explores whether AI-written articles reach fewer readers than human writing, especially through search. The short answer is that the corpus has no direct measurement of search traffic for AI articles. What it has is a hint, some engagement data that stands in for traffic, and a reason to think 'less traffic' may be the wrong thing to measure. The hint comes from the Graphite analysis, which found that the share of new web articles written by AI rose to about 50% and then stopped growing. One possible explanation is that AI articles perform badly in search, so publishers stopped making more of them. The critique of that analysis is clear that this is a guess: Graphite reported no traffic data, so a plateau caused by poor search ranking can't be told apart from newer AI models simply slipping past the detector Why did AI article share stop growing after 2025?. Detection is weak for people too. A review of 30 studies found that humans spot AI content at roughly chance rates Can people reliably spot content made by AI?. Any traffic comparison depends first on knowing which articles are AI, and that labeling is shaky.

The closest evidence about audience response comes from engagement rather than search. On Medium, posts classified as human-written averaged about 128 likes, against about 69 for posts classified as AI. Comments showed similar gaps across follower groups, yet the authors call the gap fairly small and conclude that AI content still appeals to readers Do readers engage less with AI-generated social media posts?. A 680-person experiment points the other way at first glance. AI writing tools produced more comments and higher participation, but readers rated the discussion as more generic and less authentic. That drop in perceived quality spread even to conversations among people who hadn't used the tools Do AI writing tools improve online discussion or degrade it?. Taken together, AI content can attract attention without earning the same regard.

That split suggests a different picture from 'AI gets less traffic.' One note argues that AI posts win attention by being thorough but build no lasting reputation for anyone. Over time they push out human creators while the platform keeps making money Does AI content displace human influencers on social media?. Another note gives a reason readers find AI writing distant: human writing contains a built-in appeal for the reader's attention, and AI text inherits the platform's visibility without making that appeal Does AI writing lack the internal appeal to attention that humans use?. On this view, AI content can match or beat human content on clicks while doing worse at what clicks are supposed to signal.

There is also a deeper reason search may be a weak defense. Internet-era overload was a problem of access to a fixed body of existing material, and search and curation could solve it. AI-era overload is a problem of endless generation, so ranking alone can't keep up. That argument points toward labeling where content comes from and checking it on the reader's side Why do search tools fail against AI generated content?. The volume is already large: about 35% of new websites were AI-generated or AI-assisted by mid-2025, alongside a drop in the variety of ideas How much of the internet is AI-generated now?. The line between 'AI' and 'human' articles is also blurry. Writers edited AI-suggested paragraphs only 23% of the time, and their edits left the text about 96% unchanged Do writers actually edit AI-generated text before publishing?. Much of what counts as human writing in a traffic study may be mostly AI.

If you want to dig in, the Graphite critique is the place to see exactly what data is missing. The Medium study and the 680-person experiment are the best evidence on whether readers actually respond differently to AI content. The open question is whether search traffic is even the right measure, when AI content may be winning attention while losing trust.