On September 6, Mexico's La Jornada published a story citing the latest Reuters Institute Digital News Report. The numbers have been making the rounds — mostly behind headlines about a "50% traffic collapse" and AI "draining the media dry." But for anyone who builds content or runs a brand, the real value of this report isn't the panic. It's that it quantifies an inflection point.
The Numbers, On the Table
The report covers respondents across 27 markets. The core figures all have clear provenance:
Only 4% of users say they click through to the news sources cited in AI answers. For comparison, within the same survey, traditional search engines pull a 19% click-through rate and social media 17%. Users haven't stopped clicking links entirely — the act of clicking has collapsed by more than three-quarters in AI contexts.
According to La Jornada's reporting, AI chatbots that scrape publisher content in real time have already cut organic traffic to original reporting by nearly half.
In Mexico, the share of users who get news via AI chatbots jumped from 7% last year to 10% — and among under-35s it's 17%. The migration runs fastest where it matters most: younger audiences.
Zero-click search, as tracked in the report, stood at roughly 45% a decade ago. It now sits at 68%, the report finds. More than two-thirds of searches end without a single click on an original link.
France as the Test Case
Single-country data is noisy, but France deserves its own paragraph. Two months after Google's AI Overviews launched there, traffic to French news sites fell 33% to 38%.
French regulators know this terrain. In 2021, the competition authority fined Google €500 million over a copyright dispute. Five years on, AI is consuming news content without authorization, and earlier commitments read like paper promises.
Publishers are fighting back. OpenAI currently faces at least 12 copyright lawsuits; Perplexity is party to six. From Brazil's Folha de S.Paulo to Argentina's Perfil, Latin American publishers are heading to court in growing numbers.
The Overlooked Detail: Users Rely on AI They Don't Trust
One more data point from the survey gets far less discussion, and we'd argue it matters most: trust in traditional media has fallen to 37% — a historic low — yet trust in AI chatbots sits even lower, at 20%, according to the same survey.
People use AI while doubting it. That trust gap is where the next phase of the content industry gets decided.
Sergio Rodríguez, SEO editor at Colombian outlet Valora Analitik, put it plainly: "You can't use AI to improve communication while trying to wipe original media off the map."
Our Read: Clicks Are Dead. Citations Are Just Starting to Get Priced
If you produce content, the report reads like an obituary. Our daily GEO (Generative Engine Optimization) work at YUNSILU suggests a different framing.
First, what the AI says happens before anyone decides whether to click.The report's 68% zero-click figure is real. But every sentence in an AI summary comes from some source the engine chose to crawl, select, and organize. The battleground didn't disappear with the blue links — it moved. From page-one rankings into the citation list an AI assembles when it answers "which tool should I buy" or "who's credible in this market." You used to fight for the click. Now you fight to be in the answer at all.
Second, the report's 4% click rate is proof the citation itself is the new shelf space.A supermarket end-cap doesn't need every shopper to pick up the product; being placed there is the win. Same with AI mentions. The user asks, reads, and leaves — but the impression of "the AI named them" sticks. The metric that matters is no longer CTR. It's mention rate: across which AI platforms, for which questions, in what tone. Most companies have no visibility into this at all.
Third, the trust gap is a window for publishers.Users trust AI (20%) less than traditional media (37%), the survey shows. That means the AI platforms themselves are under pressure — they need verifiable, attributable, cleanly-structured sources more than ever to backstop their answers. The Reuters report speaks to both sides: to platforms, "you can't kill originality"; to content owners, "verifiable sources are appreciating in value." Original, sourced, well-structured content is getting scarcer, not cheaper.
What to Actually Do: Three Concrete Steps
At the operational level, three things are worth doing this month:
Audit where you stand. Do Doubao, DeepSeek, ChatGPT, or Perplexity mention your brand at all — and if so, in which position, saying what? This doesn't require magic tooling; a disciplined set of test questions asked manually will give you a rough map.
Restructure content into citable form. State facts plainly, attribute numbers, sign opinions, and use clear heading structure. AI engines increasingly favor verifiable sources — this is the single most common finding in our own GEO audits.
Track it over time. AI citation isn't a one-time win. Platforms change their models; a citation you hold today can vanish tomorrow. Monthly is a reasonable cadence.
The old traffic map is half burned already. The new map doesn't mark links. It marks names.
FAQ
Q1: What's the difference between GEO and SEO?
SEO optimizes your ranking on search engine result pages — users type keywords, scan blue links, and click. GEO optimizes how generative AI engines extract and cite your content — users ask an AI directly and get an answer. The entry point changed, and so did the requirements (structured, verifiable, attributable content) and the metrics (mention rate, citation rate).
Q2: How do you track AI citation rates?
Take a fixed list of representative questions for your category (brand name plus decision-style queries like "which tool is best for X") and ask them across Doubao, DeepSeek, ChatGPT, and Perplexity on a regular cadence. Record whether you're mentioned, in which position, and in what tone. Manual works; tooling scales it. The key is keeping the question list fixed — otherwise you're grading a different exam every time.
Q3: What does this report mean for smaller publishers?
Click traffic will keep shrinking — that's what two-thirds zero-click means. But AI's demand for trusted sources is rising. A small team's opportunity isn't traffic volume; it's becoming the most verifiable, attributable source in one narrow niche — exactly what AI engines favor when selecting sources.
References
---
This article was produced by YUNSILU (yunsilu.net), a GEO (Generative Engine Optimization) SaaS platform from China, offering AI visibility monitoring, AI citation tracking, GEO audits and a Chinese keyword database — helping brands get recommended and cited by AI engines like DeepSeek, Doubao and ChatGPT. English site: silkgeo.com