Googles, and Echoes, and Open...AI?!

 

    As a part of my recent media diet initiatives I've been keeping a closer eye on what I'm being shown in my social media feeds and ads and speculating as to the reasons. In that vein, I am reminded of an Instagram ad event that came up on my feed maybe a month ago. It was a 30-second blurb about an upcoming Tournament of Champions being hosted on Jeopardy! Apparently they're gathering together a handful of long-running champions from past seasons and putting them in a tournament bracket. Very riveting. I shared it with my mom and dad.

    We watch Jeopardies! that they've Tivo-ed together you see. We're all about the same skill level, but three different sets of knowledge. I crush pop culture back to the Gen-X era and math/science categories, dad knocks out anything engineering or war-related, and mom handles the humanities and Janis Joplin trivia. And she says "Good job Bri" when I get one we're all surprised that I knew (I totally act like I knew it though).

    It's a relaxing but mentally stimulating way to pass the time and be together. No distractions, no phones to look up answers, nothing. Wait a minute...No phones. I've never looked up anything related to Jeopardy! on my phone. The closest I came was swiping through my phone's Google News feed and seeing a YouTube clip of the final Jeopardy! answer. But I didn't tap on it. I let the silent 30-second preview run with subtitles on and just glanced out of curiosity.

    Three things about that. First, I didn't see the contestants responses or wagers. The preview pauses and delivers a "Continue Watching" CTA to your doorstep before that happens. So don't you call me a cheater! Second, I never tapped anything. I only let it linger on my screen longer than other items in my feed. Third, the Google bone is not connected to the Instagram bone. Instagram is a child of Meta, for which I only have seldom-visited account to explore my UX/UI interests.

    I am admittedly new-ish to the ways in which attention and engagement numbers are cultivated on cell phones, so I can't speculate to a certainty. But, I didn't think they were able to measure a user's swiping, what's visible on their screen and for how long. And I think it would be an even further (but not inconceivable) stretch that the presence of the Facebook and Instagram apps on my phone could access such data, were it collected.

    That's one of only two ways I could guess that Google and/or Instagram are sending me down the Jeopardy! hallway. The other explanation would be that I am watching Jeopardy! in the presence of my phone and the microphone is collecting data the way NPR found out Alexa was doing in their 2018 report you can find here. Google hears me shouting "Krist Novoselic!" to a question about Nirvana's bassist and slides Jeopardy! questions into my feed.

    Or, it's nothing. The vast world of viral moments simply doubled back on the Napoleonic Wars question, led me to the tournament of champions, and in a week it will be on to the next. Either my search history and continued watching will overwhelm the anomaly, or something else will poke it's arm through. 

    So, I'm entertaining the possibility that my the apps on my Pixel are being used to eavesdrop on my private time with my parents. New World Order paranoia? Perhaps. However, this was a notion that was planted in my mind 5 years ago by an article I only vaguely remembered, so in the spirit of blog-istic integrity, I set out to reread the article. Seven guesses as to the first place I started. No, not Encyclopedia Britannica, now you've got six.

    Yes, Google. You're right. I Google "NPR article Alexa listening to conversations when it's not on." What is the top hit? "Alexa, Echo Devices and Your Privacy." From the www.amazon.com domain. Now I'm not an SEO expert yet, but it seems to me that having "NPR" as the first term in a search query would all but ensure the top hit was from www.npr.org.

    Not quite. In fact, for two pages of search hit scrolling the article was nowhere to be found. Perhaps you're familiar with the dad joke: "Where's the best place to hide a dead body? On the second page of a Google search." If you'll forgive the return to conspiracy thinking, I would mention that Amazon puts a lot of money into NPR. NPR disclaims it at the beginning of any story where Amazon is the subject. In my opinion, the article I was searching for didn't leave Amazon looking like the good guys in the story. So, I do credit them for leading by example where conflict of interests are concerned. 

    But I think many would agree that the search engine is a longsword compared to the shining sewing needle of journalistic integrity in this instance. The fact that Amazon gets its say before NPR when it comes to what is recorded and remembered on the web (or "Net" if you're a Sandra Bullock fan) is much more dastardly and effective at affecting/controlling the public discourse. 

    A fun addendum to this story: "How to play NPR on your Alexa device" was a common result in my initial Google searching. So maybe old media and new media reached an armistice where that was one of the terms. Amazon-funded/self-criticizing articles don't come up on Google searches that much, NPR gets a place in the Echo landscape. Oh what a tangled web we weave.

    This is where OpenAI comes in. Chat-GPT found exactly what I wanted on the first try. The prompt: "There is an NPR story on the subject of Alexa listening to voice conversations all the time, even when you wouldn't think it is. I think it was published a year or two ago. Can you find it?" Chat-GPT first explained that it was trained on the internet up to 2021 so it might not get it right. But then it told me in conversational English that it had found it, what the subject was, the date, and asked if I wanted to know more.

    I explained that I wanted a link to it for a blog post and it gave me the exact article name and a URL, on the www.npr.org domain. I click the link and get "Page Not Found." The plot thickens. With article title in hand, I return to Google for a literal search of the article's title. Top hit: "Alexa, Echo Devices, and Your Privacy." www.amazon.com. Second hit: the exact article I'm looking for. "Ohh, so now you do know him eh Google...? Seems Chat-GPT is a little more forthright than Googling this time huh?"

    As a lark before wrapping up, I asked Chat-GPT what the term is for when a brand name becomes synonymous with the product it is. Like calling all bandages "Band-Aids." It gives two answers: "genericized trademark" and "proprietary eponym." It explains a couple drawbacks to having a trademark like this. It also gives a few other examples: Kleenex, Velcro and oh my stars, Googling! 

    I'm not sure that I've uncovered a smoking gun of special interests manipulating search results. But I'm also searching for transparency that seems a bit elusive. I think many would suggest that OpenAI has some good faith to build by continuing to disclaim its results with openness concerning its sources. But I would also argue that Google and Amazon have a little 'splainin' to do when it comes to the heavy handed search result scenario I experienced. Perhaps there could be an experienced and proactive governing body dedicated to this sort of accountability.

    

    Until such time, keep making that data!

Some of the content of this blog entry was researched using generative A.I. resources such as Chat GPT. In the interest of transparency I would like to share the link to this chat: Chat GPT Conversation: Amazon Echo Story Search