Publishers Beware: AI is Coming for Your Search Traffic
Google is again changing how the Internet works
Sree’s newsletter is produced with Zach Peterson (@zachprague). Digimentors Tech Tip from Robert S. Anthony (@newyorkbob). Our sponsorship kit.
🗞 @Sree’s #NYTReadalong: Recent guests include Ali Velshi of MSNBC and Alex Counts of India Philanthropy Alliance. Here’s the recording of Ali’s episode; here’s Alex’s. You’ll find three years’ worth of archives at this link (we’ve been reading the paper aloud on social for 8+ years now). The Readalong is sponsored by Muck Rack. Interested in sponsorship opportunities? Email sree@digimentors.group and neil@digimentors.group.
🎯 Work with us! Our company, Digimentors, does digital and social consulting, as well as virtual/hybrid events production and all kinds of training. See our updated brochure (would love your feedback). Get in touch (no project too big or too small): sree@digimentors.group and neil@digimentors.group. If you’d rather chat, here’s my Calendly.
🤖 I’m offering workshops about AI, ChatGPT, etc. They’re 20 minutes to 3 hours long, remote or in-person. No audience is too big or too small. If you know of any opportunities to present these (customized for businesses, nonprofits, schools, etc), LMK: sree@digimentors.group. Here’s the brochure to my non-scary guide to AI: http://bit.ly/sreeai2024
***
SOCIAL MEDIA WAS A BOON FOR PUBLISHERS right up until it wasn’t. I have a feeling that we will start talking about search traffic in much the same way very soon.
As I wrote back in October 2023, the falling out between news publishers and the social media giants was severe. Buzzfeed built a business—and some significant capital investment rounds—on being a social-first publisher. Now, it’s a shell of its former self and Buzzfeed News was shuttered as the cuts started. The New York Times, Washington Post and Wall Street Journal all fully embraced the reach of the social behemoths, hiring social production teams to build content to satisfy the ever-evolving algorithmic whims of the news feed. I remember when a journalist’s social following was a key element of the recruitment/hiring process. That all feels like eons ago.
As social media platforms high-tailed it away from anything resembling news, search traffic became an essential element of news publishers’ traffic. Some estimates say the number is 40% of all traffic, but I’ve heard—anecdotally from friends in the industry—that, for some publishers, Google refers closer to 70% of their web and app traffic.
Google has told us that the future of search is not links as we have known them, but “AI overviews,” which present summaries of answers when available.
The sad irony here is that the AI-generated summary in the Google results will likely come from a model trained on news publisher content. It’s the ultimate insult.
There’s a great podcast episode from The Wall Street Journal from December 2023 with Keach Hagey and Alex Ossola that I encourage you to listen to (or read the transcript). This bit from Keach sums it up nicely:
Google can only train a model if it has data to train the model on. In fact, Search itself only works if there is something to search. And herein lies the problem. At some fundamental level, Google is not incentivized to completely destroy the open web because it would destroy the utility of the Google search itself. Now, what Google says is, "It is not our intention to stop sending traffic. We intend to build this thing in a way that will continue to send traffic to publishers."
Publishers say in meetings, that's what they're hearing privately from Google as well, but Google can't tell them how. And in the product that's out in the marketplace right now, it seems demonstrable that the links are far less prominent than they are in previous versions of Google. And not only that, but that they are not the sources of the information. They are corroboration. And I know this sounds like a small difference, but it's actually not. Because the links that are there are just extra things that might support this summary. They're not necessarily the source of the summary. And breaking that link in information is really challenging for publishers.
“Challenging” strikes me as the understatement of the century! Reliable, fact-based journalism has taken a beating over the last decade, and now publishers are on the way to becoming context offerings and little else. As Heagy noted back in December, and as you can see in the demo from Google at its I/O event last year below, it’s not hard to see how a generative AI-based results page differs from its predecessors. Combine this with Google’s ongoing experimentation with eliminating the “news” tab in traditional search results and… well… yikes!
Most Google searches are not for news-related things, but that’s not really the point. In 2022, nearly half of all newspaper ad revenue came via digital channels, but that revenue relies heavily on consistent traffic. I just don’t see a way to spin this as a positive for publishers large and small.
Of course, publishers are not the only ones in the crosshairs here. I’m sure media buyers and agency folks all over the world are clamoring to get their heads around what it all means—not to mention small businesses and content creators. For those more in-the-weeds on the issue, I found this piece on Neil Patel’s site by David Shapiro, Senior Vice President of Earned Media at NP Digital, very informative. To get a good idea of just how up-in-the-air it all is, this Eric Schwartzman piece for Fast Company provides a great rundown, including the sort of existential crisis this is causing for Google as a company built on search.
The advice I would give to publishers is this: Make your core, loyal audience a priority right now. Learn as much as you can about that audience and do everything you can to nurture and grow it. Facebook has no space for news, Instagram and Threads will do little-to-nothing to promote news content, Twitter is awful, and now Google traffic will start to decline too. Ephemeral digital traffic is being degraded as the digital world expands exponentially, and the news may be a casualty of it all.
— Sree
Twitter | Instagram | LinkedIn | YouTube | Threads
My other AI essays:
AI isn’t ready to be everywhere April 25, 2024
Don’t call them “hallucinations!” April 16, 2024
Copyright, fair use & AI March 7, 2024
The great AI acceleration Nov. 27, 2023
AI companies use the public as a laboratory Sept. 26, 2023
The liberal arts are more important than ever April 23, 2023
Education in the age of AI April 16, 2023
What AI demands of us Feb. 27, 2023
Promotional message
Be smarter at cocktail parties with THIS WEEK, THOSE BOOKS
A quick weekly newsletter and five-minute podcast that connect the dots on the week's big news story and the world of books. A literary take on the news from London-based journalist and writer, news junkie and passionate book reader Rashmee Roshan Lall. Makes you smarter, faster. More than 10,000 readers in 110 countries agree. Sign up here.
DIGIMENTORS TECH TIP | The Curtain Slowly Rises on Canon’s Flagship EOS R1 Digital Camera
By Robert S. Anthony
Each week, veteran tech journalist Bob Anthony shares a tech tip you don’t want to miss. Follow him @newyorkbob on Twitter and check out his 1.1 million followers on Pinterest!
As the old saying goes, the best things in life are worth the wait. It’s a good bet that Canon sure hopes so as the company unveils the long-awaited flagship of its mirrorless digital camera line: the Canon EOS R1.
Ever since Canon moved away from single-lens-reflex digital (DSLR) cameras, which have mirrors which flip up out of the way before a photo is taken, in favor of mirrorless cameras, which are smaller and quieter, professional photographers have been waiting for Canon to top off its mirrorless camera lineup. The Canon EOS 1D X Mark III, its last flagship DSLR, was released in 2020.
In announcing the Canon EOS R1, the company provided few details on the camera, instead highlighting capabilities for sports and wildlife photographers who need to capture fast-moving subjects like birds, basketball players and gymnasts.
For example, an Action Priority autofocus setting will allow the EOS R1 to track and maintain focus on a football player even if a defender cuts in front of him. According to Canon, deep learning technology in the EOS R1 will enable the unit to detect certain actions, like a basketball player shooting the ball, and lock the focus on that player.
While the Canon EOS R1 is overkill for casual photogs and many professionals, its announcement offers a peek into technologies which will eventually trickle down into more-affordable cameras in the future.
For example, while Canon has made no such claim yet, it’s possible that the EOS R1 will include a “global shutter,” which is a vast improvement over the image sensors in most of today’s cameras.
Most professional digital cameras let you choose between a mechanical shutter, be it embedded within the lens or inside the camera, or an electronic shutter where the image sensor does all the work. While an electronic shutter has the advantage of being silent, if the subject moves too fast you can end up with distorted images.
This is because conventional image sensors don’t capture everything they see at once—they scan the image line by line. Thus, if a subject is moving too fast, the last part of the sensor will be fractions of a second behind in capturing the image—thus creating the distortion.
With a global shutter setup, the sensor captures the entire image at once, a task that sounds really simple, but actually requires an astonishing amount of smart circuitry because of the amount of data that must be captured and stored almost instantaneously.
If history is a guide, Canon will probably supply a few hand-picked photographers with pre-production EOS R1 units for use at the 2024 Summer Olympics in Paris as it did in 2021 with the then-new EOS R3 mirrorless unit for the delayed 2020 Summer Olympics in Tokyo. For the rest of us, we’ll have to wait until at least later this year.
Canon’s press release on the EOS R1 was most notable for what it didn’t include: technical specifications, accessories and of course, price. Note that the Canon EOS R3, the next-highest model in Canon’s mirrorless camera lineup, lists at $4,999, while the Canon EOS 1D X Mark III, the top DSLR, sells for $6,499. The EOS R1 will probably carry an even-more-eye-watering price tag. Start saving now.
Did we miss anything? Make a mistake? Do you have an idea for anything we’re up to? Let’s collaborate! sree@sree.net and please connect w/ me: Twitter | Instagram | LinkedIn | YouTube / Threads
As you started me on my social media and AI journies, it probably comes as no surprise that before reading this I had just finished an outline for a class I'm teaching that will mention AI's effect on journalism and news publishing that is totally in line with everything you say here. I am generally an optimist and inside me something squeaks, "things will work out." But right now louder than that squeak is a big sadness.
I am not ready to live in this AI, climate collapse world! A very sobering post, Sree.