It’s hard to treat AI doomer stories seriously these days, but the Google Overview AI story may have some merit.
When it comes to AI stories, there is no shortage of mainstream media hyping it up as some sort of death knell to an industry or humanity itself. I’ve seen claims all over the map saying that AI is going to be the end of something. Whether it is humanity itself after some vague hand wringing, that artists are going to be out of work due to AI image generation, that lawyers were going to be replaced by AI, or that journalists themselves were going to be replaced by AI, mainstream media reporting has been all the same. That is overhyping the power and capabilities of AI and shouting “we’re doomed” in an effort to get clicks.
Those stories usually end with the realization that AI isn’t actually that powerful and that AI is in no way going to actually replace human beings in the end. At the end of the day, AI may very well be a useful tool to help make a job more efficient, but it’s not going to be the overwhelming threat that the mainstream media made it out to be.
So, when I saw media reports all last week about Google’s Overview AI that is being trialled in the US, I initially chalked it up to yet another AI doomer article that was destined to repeat that cycle. An example of this was CNN saying that publishers days would soon be coming to an end:
The A.I. doomsday clock appears ready to strike midnight for publishers.
Google on Tuesday announced that it will infuse its ubiquitous search engine with its powerful artificial intelligence model, Gemini, drawing on the rapidly advancing technology to directly answer user queries at the top of results pages. “Google will do the Googling for you,” the company explained. In other words, users will soon no longer have to click on the links displayed in search results to find the information they are seeking.
On its surface that might sound convenient, but for news publishers — many of whom are already struggling with steep traffic declines — the revamped search experience will likely cause an even further decrease in audience, potentially starving them of readers and revenue. Why spend time clicking on a link when Google has already scoured the internet and harvested the relevant information with its A.I.?
“Google will take care of the legwork,” executives said. But a lot of that legwork, of course, comes in the form of human-written articles and expertise published across the internet on blogs and media outlets, all built on a foundation of advertising support.
Google’s message was heard loud and clear. Within hours of the Mountain View announcement, the news industry began sounding the alarm.
“This will be catastrophic to our traffic, as marketed by Google to further satisfy user queries, leaving even less incentive to click through so that we can monetize our content,” Danielle Coffey, the chief executive of the News/Media Alliance, bluntly told CNN.
Just reading that, you can see why I had a hard time treating this story seriously. This especially given how mainstream media has been pretty much wrong about every other prediction about AI. If a mainstream media outlet is publishing something on how AI is going to be the end of us all, you may as well be reading the Daily Mail or some other trash tabloid source that is likely completely fabricating the stories.
When I saw the story, my immediate response is that it depends entirely on how it is implemented. Will it contain source material up front? Are general search results going to still be present? Will this appear for all web queries or a select few? Questions like that are pretty big ones when it comes to determining what kind of impact this sort of thing will have on traffic and the article in question didn’t really offer much in the way of actual indication of what it may do. Just predictions on how scary this all is and we should all be hitting the panic button.
Another angle I saw some people looking at all of this is that this may all just be one massive copyright violation and that publishers can sue Google for summarizing their content. These perspectives generally get copyright law very VERY wrong. Summarizing something is not copyright infringement because you can’t copyright facts. You can copyright the expression of those facts, but not facts themselves. So, for instance, if a news source reports, “It was a quiet evening on the highway for police officers when they saw something that shocked them. Joe Smith was driving part way on the curve and speeding at 100 MPH in a 50MPH zone. Police chased him down and pulled Smith over, arresting him for a DUI.”
Now, if someone summarized that by saying, “Police arrested Joe Smith for a DUI.”, then that summarization isn’t copyright infringement. This applies to both human’s summarizing that or AI doing the same thing. So, as far as I’m concerned, there is no strong case to be made there.
So, this leads us back to what the Overview AI implementation is actually like. Since this is being trialled in the US, I don’t exactly have an easy way of determining what it actually looks like. Luckily, American’s have already seen this and posted screenshots and descriptions of what this looks like. Here’s Arstechnica:
Google I/O has come and gone, and with it came an almost exclusive focus on AI. Part of the show was an announcement for Google Search that was so huge it was almost hard to believe: the AI-powered “Search Generative Experience (SGE)” that the company had been trialing for months is rolling out to everyone in the US. The feature, renamed “AI Overview,” is here now, and it feels like the biggest change to Google Search ever. The top of many results (especially questions) are now dominated by an AI box that scrapes the web and gives you a sometimes-correct summary without needing to click on a single result.
AI Overview is a bit different from the SGE trials that were happening. First is that AI Overview is a lot faster than SGE. For some popular queries, it seems like Google is caching the AI answer, which should help with the high cost of running generative AI. For queries with cached overviews, you’ll see the AI box load instantly, right along with the initial search results pop-in. SGE responses would come in word by word, like they are being typed by a person. When you aren’t getting a cached result, you’ll see a blank AI overview box that loads with the search page, which will say “searching” while it loads for a second or two. Other times, Google will try loading an AI Overview and fail, with the message “An AI overview is not available for this search.” (As if anyone asked.)
When Google decides you have an AI-appropriate query, it now takes a lot of scrolling to see web results. Google scrolls infinitely, so there are no “pages” anymore, but let’s consider a “page” to be a full browser viewport height: The first page is an AI overview that takes up half the screen and then another answer box extracted from some website. Page two is a “People also ask” box suggesting other queries, then one search result, then a box for videos. Page three is the bottom half of the video box, then a “Discussions and forums” section with Reddit and Quora posts. It’s not until page four and miles of scrolling that we get the traditional 10 blue links. This list isn’t even counting an ad block, which would appear first normally. I’ve yet to see an ad block and AI overview at the same time, but I’m sure that’s coming. Despite pushing AI Overviews live into production for everyone on the most premium spot on the Google Search page, Google still notes that “Generative AI is experimental.”
Google says you can’t turn off AI overviews in the main search engine. I’m still seeing the “Labs” icon in the top right, with some checkboxes for AI features, but those checkboxes are no longer respected—some queries will bring up an AI overview no matter what. What you can do is go find a new “Web” filter, which can live alongside the usual filters like “Videos,” “Images,” “Maps,” and “Shopping.” That’s right, a “Web” filter for what used to be a web search engine. Google says the Web filter can appear in the main tab bar depending on the query (when would a web filter not be appropriate?), but I’ve only ever seen it buried deep in the “More” section.
Once you do find the Web filter, the results will look like old-school Google. You get 10 blue links, and that’s it, with everything else (Google Maps, answer info boxes, etc) disabled. Sadly, unlike old-school Google, these are still the current Google web results, so they’ll be dominated by SEO sites rather than page quality.
The article shows screenshots of what the AI looks like and I can admit that the screenshots did give me pause. This is for the simple reason that I have an understanding of how search results tend to work.
When Google scours the web for results of a given query, it tries to rank those results based on relevance. These days, there are Search Engine Optimization strategies that tend to override the quality of some sources, so such efforts have contributed to the decline of quality of search results for a lot of people. Still, when your average person searches for something, they tend to click on the top couple of results whenever a search result looks good. This is usually the top couple of results. If you have to scroll for your page result, then the chances of you getting clicked on by a number of people decreases dramatically. Heaven forbid you are on the second or third page because people are much more inclined to either find something else or perform a new search instead.
So, that spot that Overview AI is taking up is generally where a lot of eyeballs would normally go to. In order to view web results, users would have to jump through additional hurdles to get there. Some might be willing to jump through those hurdles, but users are generally used to little friction when it comes to looking for something on the web. There’s definitely a good chance that the default AI response would be something that some users would wind up doing.
I can admit at this point that this puts a few additional wrinkles in dismissing this story as just another clickbait nonsense story that would disappear into the background noise within a few months.
The question for me ultimately becomes how much users would adopt the AI results over human generated research. Would users adopt a strategy of scrolling past the AI to find actual results produced by humans because the product isn’t trustworthy or will users just take the AI result as-is? If it’s the latter, then I have reason to be worried given that a huge portion of my traffic comes from Google in the first place.
Probably the one thing that is comforting on an interim basis is that this doesn’t affect all search results. When people put a query into the search bar asking questions or asking for instructions, then maybe this situation is workable for the time being. Even then, if adoption is good, Google would have reason to continue expanding this to other search types as well.
At this point, I admit that this may very well be the first AI doomer story that might actually have some merit after all. A big question in my mind at this point is if I’m overlooking something somewhere along the line. So, all this time, I’ve been proceeding with caution and giving this time to percolate in my mind. Sometimes, an answer pops out after a while that I didn’t initially think of, but nothing came to mind that would dismiss this worry outright. So, I hoped that another perspective might offer something that I didn’t think of. Recently, that perspective came and… it wasn’t what I was expecting. From TechDirt:
Is Google signaling the end of the open web? That’s some of the concern raised by its new embrace of AI. While most of the fears about AI may be overblown, this one could be legit. But it doesn’t mean that we need to accept it.
These days, there is certainly a lot of hype and nonsense about artificial intelligence and the ways that it can impact all kinds of industries and businesses. Last week at Google IO, Google made it clear that they’re moving forward with what it calls “AI overviews,” in which Google’s own Gemini AI tech will try to generate answers at the top of search pages.
All week I’ve been hearing people fretting about this, sharing some statement similar to Kevin Roose at the NY Times asking if the open web can survive such a thing.
But, not everything that people are searching for is just “an answer.” And not everything that is an answer takes into account the details, nuances, and complexities of whatever topic someone might be searching on.
There’s nothing inherent to the internet that makes the “search to get linked somewhere else” model have to make sense. Historically, that’s how things have been done. But if you could have an automated system simply give you directly what you needed at the right time, that would probably be a better solution for some subset of issues. And, if Google doesn’t do it, someone else will, and that would undermine Google’s market.
But still, it sucks.
Google’s search has increasingly become terrible. And it appears that much of that enshittification is due to (what else?) an effort to squeeze more money out of everyone, rather than providing a better service.
After some considerable thinking about this, I couldn’t really see an angle that says that this story is significantly overhyped and that things are going to turn out well. After seeing a different perspective, it seems that others who know how technology works isn’t seeing that angle that says that this story can be dismissed outright either. Instead, that other perspective suggests that people adopt the more open social media service instead en-mass which may be a bit of a tall order depending on how things turn out.
There have been some people asking me how I would react if this became a major thing and traffic really does experience a massive decline across the board. My answer to that is that I do have two YouTube channels and I have options to expand into the online streaming space (which is a bit more difficult to replicate in AI for the time being). That is ultimately how I designed the site from the very beginning. I look at every angle and come up with a whole list of ways to respond to different scenarios. That way, I can respond quickly without having to scramble for solutions. This plan has been a thing since I opened the initial YouTube channel back in 2019.
Now, fast forward 5 years later, if AI Overview ends up killing my website traffic almost entirely, I do have options to continue operating. The only sad thing in that scenario is that I won’t be writing news articles like I used to. Writing is something that has always been my passion even before I started writing news articles back in 2005. The idea of that career getting cut down fills me with great dread as I would have to say goodbye to something I’ve been comfortable doing for nearly 20 years.
Obviously, we’re not there yet. I’ll have to see how this rollout actually impacts the overall website traffic. Will users scroll past it, use a different source, or simply adopt the new AI system? That remains to be seen. At the very least, you know Freezenet, even in the worst case scenario, isn’t going to die out. There are plenty of options left to stay afloat still hanging around even if Freezenet as you know it may not continue on forever. At any rate, if you were hoping for an article that finds a way at looking at this AI development and concluding that nothing bad is going to happen, well, sorry to disappoint on that front. I’m not seeing any sure thing that suggests that for the time being.
