Let’s face it: a lot has changed in the past 18 months regarding where websites get traffic. We’re at another turning point in internet history – publishers must adapt or risk fading out entirely. I don’t want to see that happen. Newspapers and other news publishers are essential for a thriving democracy. But if the rise of AI doesn’t finally push media companies to modernize and adapt, this could be the end.
Too often, publishers’ first response is to push back against progress. Penske Media Corporation (PMC) filing an antitrust lawsuit against Google over AI Overviews is a good example. We saw a similar knee-jerk reaction in the late 1990’s and 2000’s with the rise of the internet. Now, if publishers can’t figure out how to embrace this next wave of industry change, they risk driving the final nail in their own coffins.
The Real Problem
Users want answers faster. That’s not Google’s fault — with AI Overviews, they simply designed a product to meet user expectations. Click-through rates began declining from SERPs long before AI Overviews launched, suggesting the real issue lies in shifting user behavior, not AI itself.
For example, SparkToro reported that in 2020, 64.82% of Google searches ended without a click. Over the last 12 months, the zero-click rate has averaged 62.25% — still high, though slightly improved.
Publishers need to stop approaching such issues as a publisher and start looking at problems the way Google or another tech company would: Your readers want answers quickly. What can your company do to connect headlines with readers more efficiently?
Google Discover
While SERPs may no longer be a lush garden for traffic, other search-related sources remain untapped, particularly Google Discover.
Discover provides each logged-in Search user with a personalized feed of engaging content. Until recently, it was available only on Chrome or the Google app on mobile and tablets, but in April, Google began rolling Discover out to desktop users — a major opportunity for publishers.
When I first introduced Discover strategies to newsrooms in 2022, we grew traffic 5x within the first year. Today, the potential is even greater, especially with its expansion to desktop. According to Allconnect, this rollout expanded the available U.S. audience by more than 43%. In Q1 2025, Discover’s CTR averaged 8%, compared to just 2% or less for Search, a 4x advantage).
Some argue Google Discover traffic isn’t sustainable or that its users don’t convert to subscribers. But my research shows otherwise. If an author or topic continually performs well for a publication in Google Discover, that
content is 83% more likely to generate consecutive appearances (two or more stories per week) over a six month period. Likewise, if a user consistently clicks on one publisher’s content, that publisher’s stories appear more often in that user’s feed, suggesting that subscription models can benefit from repeat engagement, especially when paired with low-threshold paywalls.
Targeting Discover isn’t necessarily difficult, but does require a defined content strategy, training, and a newsroom-wide commitment to SEO and content optimization techniques. Everyone — from editors to interns — has a role to play. The biggest challenge is often mastering new headline techniques and learning which approaches work best for each story.
Try Embracing AI
Publishers have been wary of AI for several reasons:
- Unlicensed content use: Concern that AI companies train their chatbots on publisher content without compensation.
- Copyright infringement: Fear of plagiarism or outright theft.
- Traffic declines: If readers get summaries directly from AI tools, they may not click-through.
- Loss of revenue: Fewer clicks mean fewer ad impressions and subscription opportunities.
These concerns are understandable, but the most common publishers’ response has been to block AI bots via robots.txt or at the server level using tools like Cloudflare or Akamai. After reviewing the disallow settings on top 100 U.S. publishers robots.txt files, only 15% chose not to block any AI bots, while 77% blocked at least one bot from the top four AI platforms: Anthropic, ChatGPT,
Claude, and Perplexity. If most publishers block bots, how can they know whether their concerns are valid?
Similarweb found that ChatGPT news queries surged 212% in 18 months, while Google’s news queries dropped 5%. Since AI Overviews launched, organic traffic to news sites declined 26%. Yet publishers embracing ChatGPT saw referrals skyrocket 25x, with Reuters and New York Post topping the list. In June 2025 alone, AI referrals were up 357% year-over-year, with referrals to news and media websites up 770%.
This isn’t the first time publishers panicked over new distribution channels. Similar fears arose with social and search but in both cases, the benefits ultimately outweighed the risks. The same is happening now with AI. ChatGPT responses to news-related prompts are often more useful than Google’s Top Stories results, providing more sources, more links, and clearer summaries.
Pros of AI results:
- Clearer presentation of important stories.
- Multiple subtopics with several source options.
- More links (13+ in ChatGPT vs. 4 in Google’s Top Stories).
- Links displayed in multiple locations (bullets, summary, sidebar).
- Nearly half (45%) of AI users click source links to verify answers; 29% consult experts or trusted publishers.
Cons of AI results:
- Limited to publishers that allow AI crawlers, which can also result in less source variation.
- Sometimes slightly out-of-date due to limited crawlable sources.
How to Win
Change the mindset: Stop letting fear of loss prevent growth. Think of AI like a lottery: “you can’t win if you don’t play”. User behavior has already shifted, and it will continue to evolve. If publishers continue to block AI bots they’ll never be able to see the possibilities, let alone build a strategy to monetize it.
Invest where it matters: Growth requires new thinking. Bring in fresh talent with ideas and skills beyond traditional publishing. Hiring from outside the industry may be exactly what’s needed.
Research + Test: Each topic or niche has a unique audience. A strategy that works for sports may flop for entertainment or local news. Test, iterate, fail fast, and try again. Persistence is key.
Transform: Even if publishers embrace AI and optimize for Discover, these are just first steps. Publishers must continue to evolve; reacting quickly to change and experimenting with new models. Next steps may start with pursuing licensing agreements with AI companies, but what follows is still unknown – maybe it will be AI-personalized subscription bundles, or direct syndication partnerships with AI platforms.
What’s happening now with AI isn’t like an algorithm update or new social media platform. It’s a systemic shift and the sooner that’s accepted, the better off publishers will be. Developments with AI are equivalent to the rise of the internet. This is your do-over, so look back 25 to 30 years, ask what they would have done differently, then apply those lessons today.
