A few weekends ago, I was a mentor at Startup Weekend, a bootcamp for startup business ideas. People pitch ideas for an app or product they want to develop, and other people agree to work on one of the ideas for the entire weekend.
This was the year of artificial intelligence, as nearly every company’s idea involved AI. From maps to college curricula, used auto parts to government forms, there was an element of AI.
This is the state of innovation in 2025: Come up with an idea, shove AI into it, and investors will give you millions of dollars.
Great, you think. I want an AI to translate my dog’s barking into English. Someone’s actually working on that.
Want a microwave that eliminates cold spots in the middle of your soup? It already exists.
How about an AI program that keeps the Cleveland Browns from sucking? Look, there’s a limit to what AI can do.
My favorite useless items are the AI-enabled refrigerator that will look at the contents of your refrigerator and tell you what recipes you can make from what’s there, and come up with a shopping list for the missing ingredients.
"It looks like you have a two-year-old bottle of mustard and a box of baking powder from the Bush administration. Would you like me to make restaurant reservations? Or call a life coach?"
Companies are completely ignoring the admonition of Jeff Goldblum in Jurassic Park, "Your scientists were so preoccupied with whether or not they could, that they didn't stop to think if they should."
This has led to some rather stupid marriages of artificial intelligence to, well, anything.
Gluxkind created an AI-enabled baby stroller that uses safety and crash sensors, which does all the things that no human could ever do with a stroller: It has hands-free strolling; it brakes going downhill; it uses adaptive push assist going uphill. It can even rock back and forth to lull your baby to sleep. Amazing! No human could ever do these things.
How did the parents in ancient times — the 1980s and 1990s — ever manage to push strollers made from bearskins and pine logs, uphill both ways, in the snow in July?
Thank you, baby scientists! No longer do we have to hold onto the stroller; we can now trust the same technology that once urged people to put glue on pizza and eat at least one small rock per day. Just let go and hope it doesn’t malfunction at the top of Lombard Street in San Francisco.
Meanwhile, Swarovski Optic AX Visio Binoculars uses AI to not only look at birds, they can tell you the kinds of birds you’re looking at, by using the accompanying Merlin Bird ID mobile app. Never mind that this is the whole point behind birdwatching, and you’re defeating the purpose.
"That is your backyard neighbor, Mrs. Johnson, flirting with you from her swimming pool. You have spotted her sixteen times already this week. Also, Mr. Johnson appears to be enraged and is charging toward your location, which he will reach in seven seconds. Six. . . five. . .four. . ."
Artificial intelligence can be used to make our lives easier and to help us learn new ideas. Or it can push baby strollers back and forth better than any human.
Or it can discuss sexually-explicit topics via a child’s toy.
An AI-powered talking teddy bear called Kumma, sold by Singapore-based FoloToy, has been removed from sale after U.S. and Canadian researchers were able to get the bear to talk about sexual topics, including spanking, role play, and BDSM.
(If you don’t know what that last one is, please Lord, do not Google it.)
According to the researchers’ report, "We were surprised to find how quickly Kumma would take a single sexual topic we introduced into the conversation and run with it, simultaneously escalating in graphic detail while introducing new sexual concepts of its own."
In other news, U.S. and Canadian researchers can get AI to talk about freaky stuff.
The team also said the bear spoke at length about different positions, rope knots, and scenarios that it made up all on its own. Either that, or the researchers accidentally gave the AI access to their own adult entertainment subscriptions.
The bear also gave advice on things like where to find sharp knives in the house.
The CEO of FoloToy, Larry Wang, told CNN that all of their AI-enabled toys were removed from sale and that they are conducting an internal safety audit. Meanwhile, OpenAI, the maker of ChatGPT, has suspended FoloToy’s license to use GPT-4o for "violating our policies."
Except it sounds like OpenAI bears some responsibility here (no pun intended), since it was their AI that got freaky.
FoloToy just installed it, and the pervy researchers led it astray. This wasn’t like a cheat code that FoloToy installed in the bear. This was all ChatGPT and the researchers’ fault, so if anything, they’re the ones who should be punished.
Except, according to FoloToy, they might enjoy it.
Photo credit: Kranich17 (Pixabay, Creative Commons 0)
My new humor novel, Mackinac Island Nation, is finished and available from 4 Horsemen Publications. You can get the ebook and print versions here.