My love/hate relationship with Nest
Sometimes, Nest — the Google-owned “smart” thermostat — annoys me to no end. It, per Google, attempts to “learn what temperature you like and build a schedule around yours.”
It’s primarily using machine learning to predict your habits and save you money on your heating and cooling. If you didn't know, “machine learning is a pathway to artificial intelligence. This subcategory of AI uses algorithms to automatically learn insights and recognize patterns from data, applying that learning to make increasingly better decisions.”
1: There is AI in my thermostat. That’s low key pretty wild.
2: It kind of sucks at it. It does things no person would do — change the AC to 67 and then 30 minutes later change it to 68 and then later to 72. I just want it to turn on the AC at night so it’s a bit cooler, raise the temperature during the day to 72 or 73, and then back down again at night. I do not have the technical acumen to tell you exactly what algorithms Nest uses, but I can tell you that I have been thinking lately about turning off Nest’s ability to learn — if I can even do that, I have no idea — and just set my schedule on my own.
3: I am not trying to dunk on Nest. I really like the app. It’s easy to use and well-designed. I am sure the people who worked on the predictive algorithms are smart and nice and do their best. Being able to adjust my heat or AC from my phone is an incredible luxury. And if you told someone in the 90s that their AC learns from their past behavior to know when they want it cold and when they want it hot and does so without their input, their brains would simply melt out of their ears.
I realize it’s weird to have such strong feelings about a thermostat, but it made me truly take notice just how prevalent these technologies are, right now, today.
It’s all grey for me
Spoiler alert: When you open your phone with your face, it’s facial recognition. Facial recognition is AI. The problems with some facial recognition are incredibly well documented — you can Google it — and I do not have the technical acumen to tell you about the viability of the technology going forward nor the purported improvements in accuracy nor can I recommend any constraints about its uses. There are smarter people who can answer those questions.
What I am telling you is that this stuff is here, right now. And meaningfully regulating it is nuanced. People who say “facial recognition is bad and evil and we should ban its use everywhere” and then use it to open their phones have lost the plot. I’m not saying we should conflate the use of facial recognition to open our phones with its use by a government entity or other institutions that hold power in our society — but that’s exactly how conversations around these technologies go.
It’s all or nothing, and life is rarely so black and white. Conversations about these technologies should at least be nuanced enough to help people understand what is going on in the world around them — they can decide if they care or what to do about it.
For me, I don’t really care about AI being used to flip on the AC or unlock my phone (maybe I should, and maybe my apathy is part of the problem). All these smart systems collect all our data and aggregate it for training (another topic for another day), but I also don’t really care what Google is using my historical thermostat data for.
What I do care about is how everything around me recommends stuff to me. Social media, online shopping, my thermostat — everything is a recommendation engine. “Hey, you liked ABC, you might like XYZ.”
In Quartz, a reporter said, “My TikTok For You page is lovely,” when writing — rightly — about how chaotic Twitter’s For You feed felt when it first rolled out (and depending on how many spam bots you see, maybe still feels that way). I think the distinction between a recommendation engine that “works well” and one that doesn’t is pretty fascinating. I am not criticizing the author here — I have similar feelings of elation when Netflix recommends a show I didn’t know about and subsequently loved.
I will say it again for the people in the back — I am not technical enough to know the fine details of how any of these systems collect a user’s specific historical behaviors and data and habits and patterns and aggregate that with data from millions of other people across dozens of systems, from shopping habits to geolocational data to every demographic and behavioral data point that I KNOW these companies have — in order to make recommendations for me.
But I do know it happens.
And I do feel weird about the fact that everything I consume is somehow pushed through this prism of AI-powered recommendations curated just for me. This isn’t limited to the digital world — Oracle has a case study about maximizing grocery store revenue by using AI to optimize what products go where and on what shelves in stores.
Even when the recommendation engines work well, something in the back of my mind still pauses at the idea that human existence is curated by technologies like AI. That could be a powerful force for good — I happen to be a techno-optimist — but we’d all be idiots if we didn’t recognize the possibility for things to go awry.
Smarter people might have some historical analog — 30 years ago, I imagine what newspapers and cable news in the US decided to cover and not cover — and how they covered it — could radically transform how human beings experienced and understood the world around them. I don’t dispute that there’s always been some filter through which information is passed. But with AI, it feels different. It feels more common, more frequent, more ingrained, and less noticed.
Only time will tell if it’s good or bad, but it’s here all the same.