I'm not going to sugarcoat it: it's been an awful last year for small digital publishers like me. I've spent the last 15 years building this travel blog into a trusted source of information, and in less than 12 months it has gone from a really successful business I was confident in to a ticking time bomb that I'm scared of.

And it all comes down to Big Bad Google and the rise of AI.

If you also work within digital publishing, you're probably feeling this, too. I haven't heard of many small, independent website owners who haven't felt the effects of this in the last year or so. But I realized recently that the average reader of my site or follower on social media has no idea what has been happening to us behind the scenes.

So I'm here to show you.

Let me be clear: I still LOVE what I do here. The blood, sweat, tears, and tens (hundreds?) of thousands of dollars I have poured into this site over the years have been worth it, and I'm so, so proud of what I've built. There are more than 800 posts on this blog, covering 50+ countries on 6 continents around the world. I've won awards, spoken at conferences, and been interviewed as an expert on both travel and travel blogging. But the last year has been anything but encouraging.

In the last year (2024), I have lost roughly 40% of the traffic on this site alone. My ad income (which is the main way I earn money) is down 34% year over year (though a couple months ago, it was down 42% year over year). And I'm actually considered to be one of the lucky ones, as many of my fellow bloggers have lost nearly everything in the last 12 months.

So what the hell happened? It's a mixture of AI becoming more ubiquitous and “learning” (read: stealing) from sites like mine, and Google making changes to its algorithms that actively hurt small publishers in order to boost its own earnings.

I'm not gonna get super technical with this, but I do want to give you a little insight on how this has all happened so quickly.

---

In the last handful of years, AI has evolved rapidly into the GPT (Generative Pre-trained Transformer) era, with consumer-facing tools emerging that can not only automate tasks, but also generate their own content.

The advent of “large language model” tools like all the GPTs are what is currently driving the AI Renaissance. The ChatGPTs of the world that can write essays, carry on conversations, and even generate images sound cool on the surface. But underneath, they are what may ultimately be the downfall of industries like mine.

The big question (and an ongoing legal question) is in regards to that “teaching.”

When OpenAI launched ChatGPT in late 2022, it made AI technology available to the masses. Anyone could now use AI to ask a question, write an essay, brainstorm ideas, or even plan a trip. But the output that ChatGPT and similar AI tools give has to come from somewhere.

Every iteration of OpenAI's GPT (which actually dates back to 2018) has been trained on a variety of sources or parameters. The first version of ChatGPT was trained on “570 GB of data sourced from books, Wikipedia, research articles, webtexts, websites and other forms of content and writing on the net. Approximately 300 billion words were fed into the system.”

Other companies have admitted to training their AI tools on everything from Reddit threads to YouTube videos – all content created by actual humans who never gave any of these generative AI models permission to basically scrape their content. There are many ongoing court cases against the likes of OpenAI, Meta, Perplexity AI, and others surrounding this unauthorized and uncompensated use of intellectual property.

And the AI companies have admitted that they've done (and still do) this, but use the excuse that “it would be impossible to train today’s leading AI models without using copyrighted materials.” (source)

Not a day goes by without Gmail prompting me to let its Gemini AI write an email for me, or Facebook's Meta AI providing me with irrelevant links beneath photos, or emails from companies offering to “create” me an AI best friend (yes, that's real). Meta even tried to launch “AI users” recently, though it so far hasn't gone very well.

Within the travel industry alone, we're also now seeing everything from AI chat bots offering to plan trip itineraries to actual AI travel influencers (insert the largest eyeroll possible here). All of these will take not only work but also authority away from actual human travel content creators who spend money and time actually traveling to places in order to give tips about them.

An AI chat bot has no real concept of time or distance, and certainly has no first-hand experience with travel. All it takes is a person to have one bad experience with an AI-planned trip, and they'll start mistrusting all the travel content they see online.

---

In order for AI to do what it does, it requires absolutely MASSIVE amounts of energy to power giant data centers. Data centers owned by the likes of Google, Microsoft, and Meta are not new; but they are growing exponentially with the rise of AI technology.

And with an increased demand for energy (in this case, electricity to run all those super computers) comes an increased demand for ways to generate that electricity. Even though Google has made some big promises to reduce its carbon footprint, the increase in fossil fuel usage to power data centers has meant that, in the last 5 years, Google's greenhouse gas emissions have increased by 48% – 13% in 2023 alone.

It's estimated that some of these data centers will soon be using more electricity than entire cities to power the ChatGPTs and Geminis and DALL-Es of the world. Google and other tech giants are even looking toward nuclear power to meet the energy-guzzling needs of AI.

And energy isn't the only thing that AI is guzzling. The power required to run all those servers produces an immense amount heat as a byproduct. And in order to cool the systems down, many data centers use copious amounts of water; more water than anyone initially thought, in fact.

A 2024 study found that that asking ChatGPT to write you one short email consumes the equivalent of a 16 oz bottle of water – and that doesn't take into account the amount of water used to *train* the AI to be able to answer that email.

Some data centers used a closed-loop cooling system where the water can be re-used multiple times, but there's a finite number of times the water can be recycled, and some of it is always lost to evaporation. And this is all in a day and age when scientists are predicting that global demand for fresh water will exceed supply by 40% by 2030, and Google and Microsoft are already operating data centers in areas facing water scarcity.

It's estimated that a single ChatGPT prompt uses anywhere from 10x to 25x the amount of energy/water that a “normal” Google search would use. Just some food for thought.

---

As a blogger, in order to have my content “found” on the internet, I've had to play by Google's rules. Because, like it or not, Google is the main search engine used around the world. Google processes an estimated 5.9 million searches per minute, or 8.5 billion searches per day, and most bloggers have – up until now, at least – relied on Google to send them a good chunk of their overall traffic.

And a site like mine earns traffic from Google through search engine optimization, or SEO. Search engine optimization is simple on the surface: basically, you try to answer queries with your content that people are searching for, and then use specific keywords in that content to signal to a search engine like Google that your content is relevant in answering that query. Google's algorithm then decides which content is actually “best” for any query, and places that content at the top of search results.

Now look: I'm not a Google insider, and some of the things that have happened to search in the last decade may very well be coincidental, or just the natural evolution of the internet (even if I hate it). But there are several things that certainly seem connected, and that suggest that Google maybe cares less and less about humans and good search results these days and more and more about ad dollars. (Welcome to capitalism!)

Google hired Prabhakar Raghavan (the guy who was in charge of Yahoo search from 2005-2012 – you know, when it was run into the ground) to head both Google Search and Ads in 2020, and search results have seen dramatic changes since then. (This article is a fantastic if scathing look at Raghavan, who was replaced in October 2024.)

I've known bloggers who got “hit” (i.e. negatively affected) by various Google algorithm updates over the years. But they were always the outliers and not the norm; generally, if you were publishing “good” content (i.e. written by you that was either helpful or at least entertaining in some way), then your site would usually fare okay in major updates. And if you did get hit in an update, there was usually a way you could improve your site and start gaining search traffic again.

But then came the “Helpful Content Update” in September 2023. This was not the first Helpful Content Update (HCU) Google had ever released, but it was the first that seemed to target smaller publishers across the web.

Google claimed this update was going to focus on “identifying signals of quality that correspond to helpful content” – which sounds like a good thing, right?

But what happened was many small websites and blogs suddenly saw their content falling down in rankings, or disappearing from search results altogether. At the same time, we saw an increase in sponsored content, thin content from huge brands/publishers, and outdated content from forums like Reddit, Quora, and TripAdvisor suddenly appearing at the top of search results.

Google has kept insisting that they are not slapping entire sites with an “unhelpful” qualifier, but the fact is that very few sites that were hit in the September 2023 HCU have seen any significant recovery.

SEO experts, site developers, bloggers… we've all been at a loss. Every single one of my blogger friends saw less traffic in 2024 than in 2023. And, in some cases, the losses were absolutely devastating; I'm talking up to 90% traffic (and income) loss for some sites.

And the sites that were hit the worst all had one thing in common, regardless of topic or industry: they were all independent (i.e. small) websites. (It would later come out in leaked documents that Google has been adding a “smallPersonalSite” tag to identify smaller independent websites – I wonder why?)

A second blow came in March 2024, when another core update hit even more small sites. This was around the same time that Google signed a $60 million deal with Reddit that would allow Google to train its new AI tech on Reddit posts. Suspiciously, this is also when Reddit posts starting showing near the top of every single Google search. (Coincidence? Hard to believe it was.)

In October 2024, in what was perhaps just a fluffy PR move, Google hosted several independent bloggers hit hard by the HCU and other recent updates for a “Creator Conversation Event” at Google headquarters in California. To talk and “listen,” they claimed.

There are various reports from bloggers who attended this event (read this one or this one), and they all agreed on a few things: namely that Google was trying to deny that big brands (and Reddit) have been given priority in search results, that sites hit by the HCU likely would not recover anytime soon, and that they *didn't know* why some objectively good sites were so negatively affected.

What I extrapolated from this is that Google's own engineers don't even fully understand how its algorithm works any longer, and they could not roll it back or reset it to a pre-HCU time even if they wanted to. Search has changed forever.

Not to be left in the dust of generative AI and other tech giants like Microsoft and Meta, Google introduced its own chatbot, Bard, to the world in the first half of 2023. Bard got off to a rocky start, but eventually evolved into Gemini.

But Google took its AI integration to another level in May 2024, rolling out “AI Overviews” in its search results as part of its new “Search Generative Experience (SGE).” These overviews now appear at the top of most search result pages, and summarize some of the top search results right there on the page.

The two main problems are: Google's AI cannot differentiate between accurate information and spammy misinformation (which is how for a while it was suggesting people add glue to their pizza sauce), and it's just regurgitating content it's “reading,” thereby taking traffic away from the sites it's pulling info from.

A site like mine does not make any money from someone seeing something I wrote in a Google AI Overview; I only make money (for example, from ad impressions) when someone clicks over to my site and reads the content on my page. But why would they need to click over to my page if Google is showing my content right at the top of search results?

There was plenty of uproar when AI Overviews first launched, both over the sometimes contradictory nature of answers, and the fact that it was essentially stealing content from publishers without any credit.

Google has finally started adding some citations to its AI Overviews, which is at least a step in the right direction.

But those AI Overviews coupled with the other changes Google has made in the last year have led to people just not finding content on sites like mine any longer.

Now, though, when you perform a Google search, you might see a large AI Overview section (the screen real estate this section takes up is only increasing), followed by some links to (possibly outdated) online forum posts, followed by ads from few big-brand/marketing-focused websites, followed by a slew of video links, and THEN maybe some website content related to what you were searching for. Maybe.

But you might also now find yourself returning to that Google search bar to try to re-write your search query because Google isn't really showing you what you asked for.

And this is 100% by design.

Google WANTS you to spend more time within its own environment. The more times you have to run a Google search, the more opportunity Google has to show you ads/sponsored content, and push you towards their own tools. Basically, the longer you spend within the Google ecosystem, the more money Google stands to make from you.

In 2024, Google saw record profits and revenue, despite more and more people expressing wariness over the reliance on AI in search results, and Google's search market share falling below 90% for the first time in a decade.

---

But I do ask that you consider the broader implications of a world that increasingly relies on AI. My job will not be the only one that AI comes after.

And I mean, we all remember several years ago when an early Microsoft AI chatbot turned into a Nazi on Twitter within a day, right?

AI is not – and never will be – perfect, and just blindly accepting everything that an AI tool outputs as factually correct and morally sound is a mistake.

There are plenty of positive use cases of AI, but there are plenty of tasks people are using AI for that they just don't need to.

I don't WANT to see AI write novels or create movies (have you seen this? it's terrible). I want it to transcribe a podcast to make it more accessible, and let me know if I've done the math wrong on my accounting spreadsheet. I don't need AI to build me an itinerary for a city it has never visited when there are real people out there who have been to that place who can help me instead.

And when those conversations begin, pay attention to the lawmakers who get it, and who aren't in the pockets of the mega-corporations looking to save as much money as possible by replacing humans with tech. (Then again, who am I kidding? In the US, they'll replace us all with AI in a second if it means billionaires will make more.)

Other alternatives that don't rely on AI include DuckDuckGo, Ecosia, and Brave. And Bing is an option if you're just looking for less-manipulated search results (though Bing is owned by Microsoft, and uses AI).

And if there are travel bloggers you follow regularly, maybe consider checking their site for travel info first before turning to ChatGPT or an old Reddit forum from 2015 when you're planning a trip of your own. (And if you can book a hotel or tour through a travel blogger's affiliate link, that's also super helpful, too!)

Others I know are testing putting up paywalls on their sites, running group tours, or shifting over to being travel agents instead of travel bloggers. There are lots of options, and no one knows what the best path ahead is.

What I'm NOT contemplating yet is quitting.

I am very aware that Google may continue chipping away at my traffic until I'm no longer making enough money to do this as my full-time job. And I will mourn that like the death of a loved one if/when it happens. But we're not there yet, and I plan to continue traveling and publishing new content on this site as long as I can.

My strategy isn't changing that much, though I do think I'll focus more on a mixture of content types and topics, keeping in mind that I need to write things that have a chance of catching peoples' eye on social media since they're not likely to find most of my content in search results any longer.

I don't plan to stop blogging. But I am starting to come to terms with the reality that I might, at some point in the next couple years, need to declare defeat and turn to something else as my main source of income. Many of my friends are sadly already there.

But I don't see a future where the internet doesn't exist. How people access information might change, and I certainly do not condone AI bots stealing all the work I've personally done. But I do still believe there is a demand for content like mine, if only people can find it.

So if you value content created by humans for humans like what you find on this site, please consider speaking up and showing that support in whatever ways you can.

Thanks for reading. If you got this far, you deserve a cookie.

– Amanda