On Mon, Jul 6, 2026 at 9:42 PM Pascale Camus-Walter via Wikimedia-l <[email protected]> wrote:
> Hypothesis 6: is ChatGPT the culprit? > Conclusion: it does not appear as though standalone LLMs have been the > primary cause of declines in Wikipedia’s traffic. I don't think most folks who attribute traffic decline to AI would situate that chiefly with direct chatbot interactions. However, search itself is being reshaped to an extent that you cannot really even just look at where Wikipedia sits in rankings by itself as a measure of search performance. If you search for something like "causes of World War II" on Google, Wikipedia ranks #1 -- but it does so right below the AI summary. And, surprise surprise, those summaries do cost clicks. See e.g. this Pew study from a year ago: https://www.pewresearch.org/short-reads/2025/07/22/google-users-are-less-likely-to-click-on-links-when-an-ai-summary-appears-in-the-results/ Google and other search engines have been continuously optimizing those summaries since, so I would expect the effect size to increase over time. I personally start many of my searches on g.ai (their more robust AI Mode) directly; a lot of what we used to call "search" is IMO headed that way. Attribution of specific traffic patterns is always tricky, of course. But AI summaries are a lot "closer to the metal" when it comes to organic traffic drivers than ChatGPT. Warmly, Erik _______________________________________________ Wikimedia-l mailing list -- [email protected], guidelines at: https://meta.wikimedia.org/wiki/Mailing_lists/Guidelines and https://meta.wikimedia.org/wiki/Wikimedia-l Public archives at https://lists.wikimedia.org/hyperkitty/list/[email protected]/message/55OEXKM2P6HSSHU55IGMQEJCZ2THLMEM/ To unsubscribe send an email to [email protected]
