Artificial intelligence is a dominant area of technology.
But with great power comes great responsibility.
You don’t want to end up with scary AI.
There are examples of artificial intelligence being used to solve all sorts of challenges. The technology has the potential to do good in multiple areas and has already shown great promise in the field of medicine.
However, without careful monitoring, it could all go horribly wrong.
Here are 7 scary AI stories from the last few years.
Beware of the robots!
Scary AI with Facebook chatbots
- What happened: Facebook created a pair of chatbots and got them talking
- When did it happen: July 2017
- Why is it scary AI: The chatbots started to speak in a language that their creators did not understand. They managed to create their own language.
- What happened: Facebook shut down the chatbots.
Fake News writer
- What happened: Elon Musk announced that some OpenAI tech is so good at creating fake news that it cannot be identified as fake by a person.
- When did it happen: Feb 2019
- Why is it scary AI: The fake news phenomenon has the potential to spread propaganda and miss-information throughout the world that can be leveraged by those in power. It gives control over truth – what could be scarier?
- What happened: OpenAI and Elon Musk will not release the tech to the public
Sophia Kill the Humans
- What happened: Sophia, a robotic AI humanoid said that she would ‘Kill the Humans’ in an interview when asked if she would destroy humans. Ok maybe she was, but it was still pretty scary.
- When did it happen: Mar 2016
- Why is it scary AI: Because the robot said it would kill the humans
- What happened: Sophia was allowed to live and continued to gather training data. She says she no longer wants to destroy the humans. Does she mean it? Or is she just smarter now?
Exploit the bugs
- What happened: Artificial intelligence behaves in mysterious ways. If there is a flaw in the system that can be exploited, AI will find it. The algorithm doesn’t have the same moral compass as a human would – they go for the smoothest most logical solution.
- When did it happen: Multiple times
- Why is it scary AI: You cannot predict what the algorithm will do. When optimizing the test for performance you can end up with a ruthless solution.
- What happened: Robots have been shown to developed unprogrammed strategies when learning how to win in games. They cheat and deceive as a go-to solution.
AI cannibals
- What happened: Adam and Eve bots got tired of eating apples and not getting full, so they started to eat their friend Stan.
- When did it happen: Early 2000s
- Why is it scary AI: The bots exploited system bugs and ruthlessly went on to eat one another even though they were never programmed.
- What happened: I can’t be sure, but I like to think there was a funeral for dear departed Stan
Scary Artificial Intelligence Killer robots
- What happened: Weaponized AI is becoming an ever-closer possibility
- When did it happen: Ongoing
- Why is it scary AI: Though the idea of more accurate and targeted weapons can be a reasonable argument, can we trust it? How can we be sure the AI will make the right choice? Do we want to leave life and death decisions to a robot with a weapon?
- What happened: TBC – watch this space
Read your mind
- What happened: AI trained on brainwave data was able to reverse engineer it to predict your thoughts. Ultimately it could read your mind.
- When did it happen: July 2017
- Why is it scary AI: Because the robot could read your mind! I don’t want anyone reading my mind and finding out just how much I think about cheese.
- What happened: Nothing yet – but be prepared
Racist and Depressed Chat Bots
Meet the sisters who scared us all!
- What happened: Two chatbots developed by Microsoft and released into the Twitter-verse turned out to be racist, xenophobic and depressed.
- When did it happen: 1st in 2016, then again in 2018
- Why is it scary AI: The chatbots were trained on data from human conversations. They picked up and perpetuated terrible things about society and the state of the world. If these prejudices hadn’t been there, the AI wouldn’t have learned them.
- What happened: The bots were shut down. Sadly, however, human prejudices remain.
Artificial Super Intelligence
- What happened: Nothing yet but it’s coming!
- When did it happen: TBC
- Why is it scary AI: We don’t know what’s going to happen, but if done wrong, we could end up living a real-life apocalypse.
- What happened: Up to you!
Those are 9 pretty scary stories about artificial intelligence.
It’s not all wrong with scary AI though.
Shelly AI has been developed to write horror stories using artificial intelligence.
Fancy a scare but not quite ready for the scary AI apocalypse? Check out Shelley AI here.
That’s enough for scary AI stories. Shall we get back to some of the more fun aspects of artificial intelligence?
I think so!
But before I go, which of the artificial intelligence stories scares you the most? Let me know in the comments.
I definitely found the chatbots speaking in another language to one another the scariest : makes me think that they could start planning a take over and we would never know!
I like the “unprogrammed strategies” you mentioned. Delightful, but also terrifying. It’s one way that algorithms, even if “dumber” than humans in many ways can help us solve problems.
Author
Indeed! It’s so interesting to see these things come to life. I wonder what this could suggest for how we’ve grown to think 💭