The internet searches I made during my miscarriage are still on a Google server

I often think of internet searches as a kind of direct thing. I type in a term, press enter and the page populates with potential answers. But searching has always been a thing more concerned with wandering, than finding. Which makes sense. It would be called “finding” otherwise. Which kind of reminds me of a joke my dad said the few times we went fishing and I got frustrated about not catching a fish, 

“Well, it’s not called catching, it’s called fishing.”

Anyways. 

I know I wander when I search for my lost keys. But there’s a lot of round about wandering going on when I type something into Google too. Automated computer programs called crawlers search the web for information, that information is categorized, ranked and stored in databases. When you or I enter a search term into a Google bar, those databases are used to give us answers based on ranking and relevancy. 

Google doesn’t just store websites. It also stores searches. Those searches are used to customize ad and search results for each user. Searches made before 2018 can be stored indefinitely. Google says it now auto deletes searches from their databases after 18 months. That is still a long time, and it means that every day you are on the internet searching that 18 months is starting over. 

Does that matter?

I mean…maybe not all who wander are lost, but most people who are searching are looking for pretty mundane things. The object of a search - in real life and online - is generally not interesting. Shoes, keys, the best sandwich within ten miles. 

But sometimes searches are full of hope or anguish or both. My hopeful, anguished offline searches only continue to exist in my head. But my hopeful, anguished online searches still exist on a server somewhere. And I find that very unsettling. 

Like, somewhere there is a server holding the searches I made when I was starting to miscarry my third pregnancy, but couldn’t get into a doctor.

“Can you bleed a lot and keep a pregnancy?”

“Are blood clots normal when you’re pregnant?”

“Do ERs do ultrasounds if you think you’re miscarrying but can’t get into the doctor?”

“How much do ultrasounds cost at the ER?”

“How long do miscarriages take?”

“What does an embryo look like at 11 weeks?”

“How much bleeding is normal during a miscarriage?”

“How do you know if there’s an embryo in your blood?”

“Is it normal to be depressed three weeks after a miscarriage?”

Those searches were used to serve me ads about infertility treatments, depression medication, self-care items and other things marketing execs have determined miscarrying women might spend money on. 

I don’t know how often the marketing worked and I don’t know how long the “post-miscarriage ad suite” was served to me. I have no idea how my search results in the weeks and months and years after those miscarriage searches changed the way I decided to live my life. 

Isn’t that…wild?

Every time I’ve turned to Search to gain some knowledge, I’ve also suffered extraction. And I’ve never received compensation for the things pulled from me. And I am never sure how they’re being used to change my future, so I do not know what the extraction kept from me. (Oh my gosh, please read The Age of Surveillance Capitalism for a DEEP dive on this.)

Now, with the introduction of generative AI text models, it’s the future of Search that seems to be up in the air. Google results can’t compete. 

ChatGPT is a language model that can generate human-sounding text. It’s trained on words from all over the internet. Basically, it’s a chatbot that uses machine learning to answer questions posed by people like you and me. Ask Jeeves - but cooler. My kids always make fun of me because I phrase my google searches like questions. But now that’s how a lot of people are searching. 100 million people are already asking ChatGPT questions and getting answers every month.

Some people think chatbots will put writers out of work. This experiment I did with ChatGPT kind of blew my mind. 

I don’t know what AI text engines mean for writers. I do think AI chatbots are interesting tools. I think ChatGPT’s best use has something to do with the future of internet search. That’s definitely why Google is afraid of it. And that’s definitely why everyone is rushing to market with their own AI chatbots - they want to commodify AI language models. Search is one of the easiest ways to do that because everyday people use Search every day. Its potential market is the world.

Search makes the internet work for people who don’t know how the internet works. But Search doesn’t really work anymore. All Google Search gets me these days are ads followed by pretty poor results. At least in the years before this, I got pertinent information in exchange for the extraction. Search needs some sort of a new future. And I do think we should be incredibly thoughtful about what that future looks like, because Search matters. 

Let’s look at it from the kitchen. Search connects the home to the rest of the world. Using Search and the infrastructure built around it, caretakers can sit at their kitchen tables and enter the public square. Search helps reveal that the border between the domestic space and the public space isn’t a natural formation, it’s a construction.

I am a product of Search, I guess, in more ways than one.

When I could not afford childcare to go back to college, I used Search to educate myself. Search delivered academic journals and university press books and ancient poems and modern ideas. I’ve found voices that are traditionally silenced because of Search. And I found you (and you found me!) because of Search. My writing career only exists because of Search. I am a product of Search, I guess, in more ways than one.

It’s not just my education and work that depends on Search. I use Search to find my kids’ doctors, recipes, and see online museum exhibits from museums in countries I’ll never get to visit. There are hard nights when it is too late to call anyone, but it is never too late to Search.

On those nights, Search has been pretty close to an answered prayer. When one of my kids is sick and I don’t know if their fever is too high or their breathing too shallow, Search has helped me know what to do. 

Companies like Google have taken every query and every website visit and turned it into data. That data has been used to sell me things and nudge my behavior. (Which I guess is kind of how many religions think prayer is used by whatever deity islistening…but it’s not really how I see it.) Some of that behavior was nudged after huge events like miscarriages, but lots of it has been nudged using my everyday searches. I think I notice those nudges the least, and so they probably move me the most. 

Search has enabled lots of bad searching, too. Of course. Not every query typed into Search is looking for academic articles or parental advice or “the best jeans for women with huge bums and slightly smaller waists.” (Ahem, still looking for the right answer to that last one. Even Search couldn’t deliver!) Some searches are bad. But that was true before the internet, and will be true after the internet. (Yes, I think there’s an after. No, we don’t have time to get into all of that right now.)

While everyone is asking who will win the AI chatbot fight, I have different questions. And Search can’t answer them. At least not yet. 

Can Search be reimagined as something that exists to clarify instead of extract?  And if an AI tool could deliver that change, would any of the people developing the technology actually use it that way?

When it comes to generative AI of all sorts, there are both ethical and practical questions that still need to be addressed. In his excellent article about Bing’s new AI chatbot, Casey Newton from Platformer gets into some of them,

Of course, there are plenty of serious questions about the broader implications of these tools becoming widely available. Microsoft spent a good deal of time Tuesday explaining its approach to making the new Bing safe, which largely involved describing org charts accompanied by arcane diagrams. Ultimately nobody knows how safe it is to put AI tools like these into the hands of billions; to some degree everyone here is just hoping for the best.

There are also the regulatory concerns, which came up less often but seem just as relevant to me. I’ve written before that publishers seem likely to sue over the way that their journalism was scraped to train the models on which tools like these run; I wouldn’t be surprised if at least four European countries started drafting lawsuits against Bing today based solely on what they could see in Microsoft’s screenshots.

Like a lot of tools, AI chatbots can be used for very bad things. They are only as good as the data sets they’re trained on. Those data sets come from real writers, how will they be compensated? As it stands, generative AI is built through extraction. I certainly haven’t seen a check or citations for the way my writing has been used in the dataset. And even with training, a lot of the information ChatGPT gives is incorrect. 

Still.

The potential for a better Search through AI chatbots is there. Especially if, like Neeva AI, it gives its answers with citations! (I just signed up to test out Neeva - a browser that is privacy-rich and ad-free. After a month or two, I’ll let you know what I think.) And a better Search would really make things…well…better, at the kitchen table and away from it.