Before we jump in, note that this week's newsletter is going to be spicy. It's going to be a different vibe from the first two, so bear with me.
Let's first talk about AIOs.
AIO is arguably the pinnacle of AI slop
This week, Lily Ray coined the term "[AI-splaining](https://x.com/lilyraynyc/status/1913989383062962364)" to describe when a searcher types a made-up phrase (no matter how silly) and adds "meaning" to the end of it, AIO will make up something.
I've also tweeted a couple of instances when AIOs were clearly providing the wrong answers and citing pages that provided the right answer.
I truly worry about how AIOs may be providing false information and the number of people out there who will assume that whatever AIO says is true. What also bothers me as an SEO is that we were taught about EEAT but AIO is anything but trustworthy.
With all that said…
Google isn't going to fix AIOs
I'll explain why in a bit.
Let's go back in time, before the internet.
If we wanted to look for information, we went to the local library. If we wanted to find a book about a specific topic, there was a computer that had a database of all the books. And before that, we had to flip through the library index cards (aka card catalog). Take it one step back further and you could say the earliest "search engines" were librarians.
Now let's fast forward to a time right after the birth of the internet. There were multiple search engines, but given PageRank (amongst other things), Google eventually came out at the top. Search behavior was originally dispersed across different search engines and eventually converged so that most searches were happening on Google.
It was that way for over 20 years, until ChatGPT changed the game.
That left Google with no choice but to launch Bard ([February 2023](https://blog.google/technology/ai/bard-google-ai-search-updates/)), which later rebranded to Gemini. Search Generative Experience, now AI Overviews, launches shortly after ([May 2023](https://blog.google/products/search/generative-ai-search/)), and it feels like it was meant to be a hybrid solution.
SGE/AIO was a way to keep people on Google search while getting the benefits of LLM. That's probably why they initially called it "Search Generative Experience" so that more people (namely those who weren't using ChatGPT) could experience generative AI within a search engine. It was a way to increase people's comfort level with LLMs that were meant to make searching for things faster and easier.
More and more, I'm convinced that AIOs are no longer a "hybrid" solution but an interim one. If Google were to acknowledge (even if it's just internally) that humans soon will not need to go to a search engine, then Google will refocus most of their resources on Gemini.
Since the creation of search engines, humans were the ones going to a search engine to find information. If the act of searching for information pivots away from the direct usage of search engines, Google needs to pivot.
Which means there isn't much incentive for them to fix AIOs. Ever.
The extinction of search engines
You might be tired of hearing this, but the future of search is fundamentally evolving as I type this. As search behavior is evolving and LLM technology improving, there isn't a strong reason for users to use search engines if LLMs can do the job — whether it's quick informational searches, deep research, or creating AI agents to do specific tasks.
I argued with Gemini for 20 minutes about this yesterday and I (or should I say "we?") cannot think of a good reason why, in the future, humans will prefer to use search engines as we know it.
In a world where time is money, even Gemini says that a core principle behind how it was designed is to "help you find the information you need as quickly and easily as possible." — so, remind me why do we need search engines?
Right now, we're in this "transition" stage — we might use search engines, LLMs, or specific social media platforms (YouTube, Reddit, etc.) depending on our desired outcome.
But the day will come when there isn't a need for us to search via a search engine. LLMs can act as a proxy for us. LLMs become the new entry point whenever we need to do searches. They'll become our personal web crawler and information retriever. Although there might still be a place for the web index, I'm keen to see if ChatGPT or other players are able to change the way the world's information is organized.
*I mean, to this day, what percent of the population still relies on librarians or the library card index? Probably just a handful of people.*
**Thought prompt**: Given what you know about search behavior today, how do you think it will change? Is there still a place for search engines in the next 5-10 years? And, should you be trying to optimize for AIOs if it likes to spout nonsense?
Yours truly, Debbie Chew
*Originally published on [Substack](https://seophilosophy.substack.com/p/seo-philosophy-3-whats-the-meaning).*