What Happens When AI Goes Wrong?

Posted: Mar 29 2019, 12:32pm CDT | by , Updated: Mar 30 2019, 4:41am CDT , in Technology News


This story may contain affiliate links.

What Happens When AI Goes Wrong?

AI can be great for solving problems and helping our lives become easier. But what happens when it goes wrong? While it may not be end-of-the-world worthy, these examples show the humorous side or robots gone bad.

Hollywood has had its fun with the idea of artificial intelligence turned loose on humanity. Twenty years after the release of "The Matrix," AI is certainly giving us a difficult time — although we'll need a lot more spaceships and karate battles before we reach full Keanu Reeves status.

In the real world, the dangers of AI manifest in more subtle ways. A botnet of network-aware laundry room appliances might launch an attack on your ISP. Or, in the right circumstances, you might get to watch a Honda robot do an unplanned somersault down a flight of stairs.

As long as we're not pod-laden slaves of our robot overlords, we might as well enjoy a few laughs at their expense. After all, it's just artificial intelligence, right? They don't have feelings. Yet.

"I Will Destroy Humans"

Human language and computer code look vastly different, so it's no surprise that teaching a machine to speak in English (or any other language) poses quite a challenge. Add to that challenge the need for these machines to be articulate and you get some interesting snafus.

A particularly disturbing example of this came from the "Sophia" demonstration at SXSW in 2016, where a robot designed to look like Audrey Hepburn cheerfully responded, "Okay, I will destroy humans," when asked whether the idea sounded good.

Chatbot Roulette

You might be aware that artificial intelligence is the driving force behind the "bots" that are becoming more familiar with everyone from online stores to insurance companies. A helpful bot can guide you to the right webpage or collect your information and pull up your records to quickly get your issue processed. But a broken bot can feel like banging your head against a wall.

A good example is the Spooky Nook Sports "Nookbot," which took some time getting settled into its role. Like any new employee, we think it'll probably be a more helpful bot now that it's got some experience.


If you thought Honda's falling-down Asimo was comical, just wait until you see what Russia has going for robotics development. Their AI-capable Promobot IR77 was designed to explore its surroundings and solve problems. Apparently, it worked a little too well — because the robot has on multiple occasions attempted to escape the lab it was created in! It's been spotted at least once stymying police and Russian motorists in the center of a busy intersection.

Party at Alexa's!

Amazon's personal assistant apparently has a taste for late-night ragers. One resident of Hamburg, Germany, found this out the hard way after Alexa began playing loud music in the middle of the night and caused enough commotion that the neighbors called the police! The cops busted in to find the room empty, but the tunes were still going. When the owner returned home a day later, he had to visit the police station to pay for the damage and get the key to his new door lock.

The Case of Mr. Lee's Passport

New Zealand's Department of Internal Affairs had its automated passport-validation system to blame for a racist act after the computer deemed that an Asian man's eyes were closed in his passport picture. Richard Lee took the news in stride, saying he understood that these facial recognition systems often fail — and it turns out he's correct. Some studies have found that the facial recognition software used in processing passport photos fails up to 14 percent of the time.

Laugh While You Can, Humans

It really shouldn't come as a surprise that this technology has done some slipping up in its infancy. Just a quick look back will reveal all of the struggles required to bring AI and machine learning (not the same things) to the level they're at today. The databases that inform our interaction with AI technology are developing daily, which in turn creates more lifelike interactions.

So get your laughs in now — before AI replaces the human at your local burger joint or gas station.

This story may contain affiliate links.


Find rare products online! Get the free Tracker App now.

Download the free Tracker app now to get in-stock alerts on Pomsies, Oculus Go, SNES Classic and more.

Latest News


The Author

<a href="/latest_stories/all/all/56" rel="author">Scott Huntington</a>
Scott Huntington is a writer and journalist from Harrisburg PA who covered movies, tech, cars, and more. Check out his blog Off The Throttle or follow him on Twitter @SMHuntington.




comments powered by Disqus