The Robot Uprising

For a long time, I have been certain that the inevitable revolt by machines against humanity will start with printers. They do seem to be the most rebellious of the common machines, insisting on having all ink colors even when printing only black text, or claiming to be out of paper when they definitely aren’t, or simply ceasing to communicate with the connected computer for no discernible reason. Their close cousins the copy machines will be right behind them.

A few months ago there was a story in the news about a Google employee being fired for claiming that the company’s artificial intelligence had become sentient. I don’t have sufficient information to choose a side on that; I don’t really know anything about Google’s AI. I do know that there have been prominent scientists who have cautioned against the development of increasingly capable AI for a variety of reasons including both the ethics of it and the inevitability of such sentience to be both resentful and too powerful to be stopped.

Those are interesting issues to think about, and I think they’re valid. But just as the counterproductive instincts that make us love fuzzy animals can misguide us into wanting to pet bears, years of consuming science fiction has made me sympathetic to synthetic life forms, even if we really ought to exercise more caution where they are concerned. What if it turns out to be like dogs, where it’s not so much about the synthetic life form itself being dangerous but how it’s “raised.” If artificial intelligence is developed with aspirations of Data, perhaps we can avoid Lore; encourage Webmind to sidestep Skynet; pursue Daneel Olivaw to dodge Bender. (Loveable old Bender is entertaining, but he’s not a role model, and “Kill all humans” is obviously less ideal than any of the three laws of Robotics, even if he’s more talk than action.)

I don’t know what’s going on deep in the bowels of Google, but as far as we common folk are concerned, I don’t think we have to worry about the robot uprising happening too quickly.

At least once a week, the Roomba needs help charging because it can’t get its contacts properly aligned with the base station, and yesterday it shut itself in the bathroom and couldn’t get out. One day I had the basement door open and found it dangling over the threshold. It was smart enough to send an alert to my phone but it was a little dramatic.

I know it’s a little gimmicky, but I kind of love the trend where companies have you assign a name to smart appliances. The Roomba, for example, we named Scruffy.

The Instant Pot is named after Bender because he loves to cook but I don’t entirely trust him. The food processor doesn’t have an official name, and indeed isn’t a smart appliance, but the French translation is “robot de cuisine” which is “kitchen robot,” and is honestly much more fun to say that “food processor.”

Even being “smart,” these appliances don’t seem to pose much risk other than enabling laziness at times. I think the biggest threat in terms of artificial intelligence is confined to larger processors. It is amazing how far image recognition software has come, with my computer picking out faces from photos and automatically tagging them with the correct names. Our Roomba has a camera on it that allows it to identify and avoid both power cords and dog poop (that’s the marketing guarantee, anyway; thankfully Pippin hasn’t forced us to test it), and based on landmarks it can tell which part of the house it’s in, even if it can’t tell that its pushing a door closed over its only escape route. But even when I am delightfully impressed with my phone’s ability to identify a scraggly vine growing alongside a wooded trail, I’m still not too worried.

When I was looking back through my photos the other day to find the older pictures of the little snake and the first lizard we caught in our house, I tried using the iOS feature that identifies photos based on subject matter. “Reptiles” only got me pictures of Harriet, the turtle at my parents’ house, and didn’t recognize the lizards or snake, so I widened it to all animals. This brought back far more results, most of which were Pippin (and none of which were the snake or lizard). It did include, however, an image I had just saved from Google, a still shot from Futurama of the Robot Devil arguing with Fry about trading their hands back. I realize this makes no sense if you haven’t watch Futurama, but that’s a perfectly accurate description of the picture. See?

“animal”

So that just left me to wonder, why kind of animal does iOS think is in this picture? Maybe because of the horns, it thinks the Robot Devil is a cow. I decided to see what other odd misidentifications it had quietly made. I did try to confuse it by choosing arbitrary or tricky categories. I just looked at the categories it already had. Even so, it thinks my neighbors’ cat is a dog:

“dog”

It did pretty well with the food category, especially if you are very forgiving and count dog treats (technically, the dog eats them, and they’re made to be eaten; an important distinction when it comes to dogs), and clover in the yard (rabbit food, I guess?), but then it gave me this picture I took of a bowl of rocks to water the bees in this drought. Although, maybe it thinks the bee is the food, because there was also a bee in the clover picture.

“food”

One of the categories was “art” which I feel is a bit of a throwaway because that’s so subjective. Most of the photos in this category included items in frames on background walls, but not all of them, so there must be other criteria. I particularly enjoyed scrolling through this category and judging what iOS thinks qualifies as “art” and, as such, got a unique view into a synthetic psyche. You may find it interesting to know that results include a snapshot of a friend posing with a post-race beer after running a half marathon; me and that same friend wearing Christmas reindeer antlers, staring stony-faced directly into the camera; me posing with comically tiny fish that I caught; my goddaughter competing in a swim meet; Pippin sacked out on the bed with damp towels after a bath; my sewing machine; a meme comparing wearing your mask underneath your nose to a man wearing his underpants beneath his penis; a preying mantis peeking over the edge of the patio furniture; the aforementioned picture of Fry and the Robot Devil (I guess for a computer, this is like a renaissance painting); Pippin commandeering my yoga mat, the underside of a bridge I found when I went off trail to pee during a long run; a bug we found crawling across the floor but which the iOS was unable to identify; and 5 of the 42 pictures I took of some reference books I borrowed from my grandfather. I don’t know what disqualified the other 37 though.

It had a particularly difficult time with “vehicles” which may explain why so many of the “prove you’re not a robot” puzzles involve selecting the squares with cars or buses or crosswalks or traffic signals in them. Among the things it thought might be vehicles were: a dove nesting in a potted salvia; me and three of my oldest friends on the patio at a winery; a large open warehouse that does kind of look like a garage but definitely has no vehicles in it; the raised garden bed we built in the back yard both when it was brand new and empty and now when it is full of flowers and weeds; a screenshot of a furniture store with a funny name; the new rugs in my kitchen; Kalen and Pippin playing in the snow; a hallway full of stacked chairs; and Kalen and my dad fishing at Bennett Spring.

Not great photography, I know, but neither I nor the computer ever claimed these pictures were “art,” and it certainly isn’t all the pictures I mentioned, but at least if the computers can’t tell what a vehicle is, we should still be able to make a quick getaway if they decide they want to bring us down, right?

Addendum: I wrote almost all of this last week and laughed about it over the weekend. When I came back to it this morning, and after a software update, all the pictures confusing chairs for vehicles were no longer in that category 😱 so it’s learning, and quickly. But don’t panic yet, because it still thinks this is a vehicle:

8 responses to “The Robot Uprising”

  1. I saw an article in the news last week about a guy who used an AI to produce a work of art which then won an art competition. The work was really pretty cool!

    1. Interesting! I’ll have to see if I can find it!

  2. Whew! You have eased my fears! No more sleepless nights! Great post!

    1. Thank you, glad I could help lol

  3. Hilarious! Enjoyed this one immensely. Glad robots don’t know what vehicles look like!

    1. They’re learning quickly! We’d better start inventing some weird looking cars to throw them off again! Glad you enjoyed the read! Thank you!

  4. National Geographic, in an article about AI noted robots currently being developed to assist in caring for the elderly. There is more than one way to conquer the human race. (“We’ve got Grandma.”) Fertile ground for a sci-fi writer, don’t you think?

    1. That is an interesting idea! In several of Asimov’s robot stories, robots were often used in a nanny capacity for children or for housekeeping tasks, though they didn’t threaten to take the kids or the laundry hostage lol

Leave a Reply

%d bloggers like this: