We often talk about how beneficial and important AI is for healthcare, finance, security and other business areas. Now let’s chill for a bit.
Today’s post, created by Magora, is a collection of the funniest, weirdest and most unbelievable cases of AI usage from the last few years. Enjoy!
Singing robots, beer crafting and a new Harry Potter book
1. Sofia the Robot now can sing
Sofia is the first android ever to be granted citizenship (United Arab Emirates) and a credit card. It was created by David Henson, an MIT PhD and CEO of Henson Robotics.
- The robot often participates in interviews on TV and shares her opinion on different matters.
- For example, in an interview she gave to CNBC in 2016, she claimed that she wanted to go to school, study business and help humanity with the integration of new technologies. However, she also said that she wants to destroy humans.
Last year she was invited to appear on The Tonight Show Starring Jimmy Fallon. There she made jokes, sang karaoke and seemed to have overgrown her ambitions.
2. Robot for beer crafting
Having a sophisticated sense of taste is crucial for crafting beer. The robots of today are still unable to admire the taste and flavour, but what they can do is collect feedback from humans.
IntelligentX Brewing Co., headquartered in London, was the first brewer ever to introduce beer made with the help of artificial intelligence. Their chatbot would ask customers about their preferences and rate them from 1 to 10. It would then interpret the info and pass it to the professional brewers.
3. Neural net writes fanfiction
“Harry Potter and the portrait of what looked like a large pile of ash” is the title of a Potter novel written by AI.
- Botnik Studios, a comic writing enterprise that uses machines to make content, created an algorithm that learned to write based on J.K. Rowling’s books. And while the three chapters that you can get acquainted with online probably won't bring you as much pleasure as the original, they are key to understanding the logic of AI.
4. Microsoft’s racist AI chatbot
In 2016, Microsoft began experimenting with a chatbot for conversation and speech recognition for Twitter. They called it Tay.
- As soon as it was launched, users began to tweet all sorts of misogynistic and racist comments at it. As a result, the chatbot could overpower Hitler himself in how categorical its opinions were.
- Being effectively a parrot who can process natural language, Tay has absorbed the worst of internet trends into its personality.
5. AI that creates art
The founders of this project were inspired by the way our brain works when creating something like it. On this website, you can turn any of your photos into a piece of art inspired by some famous artists.
- Of course, the algorithm sometimes fails, but overall the results are impressive. All you need to do is to upload your photo and choose the style set you like from the predefined options available.
6. AI that dreams
Do you want to know what the computer really sees when looking at an image? What the computer’s dreams would be like?
Check out the Deep Dream Generator project by Alexander Mordvintsev. The Google developer tried to understand what deep neural systems see when they look at a picture. In the process of his research, he has collected a bunch of tools that have proven AI can generate fresh visual content. Whether it is real art or not, you can judge for yourself.
7. German Alexa likes to party
How would you feel if you come back to your apartment after a trip to find that your key no longer fits the keyhole? That’s exactly what happened to one German owner of an Amazon Alexa device.
- He had to go to the police station to retrieve the keys and it was then that he found out what had happened – as a matter of fact, Alexa decided in his absence to play some music and increased the volume to the max. At 2 a.m. the neighbours were banging on the door and eventually called the police to a completely deserted apartment.
8. Machines judge humans in a beauty contest
In 2016, the first and most likely the last AI-based beauty contest in history took place. AI was supposed to judge the participants online, based on the symmetry of their faces, wrinkles, blemishes and so on. No glasses, beards or makeup were allowed. What was an interesting discovery was that AI refused to consider beautiful any participants with darker skin.
9. Fake news generation with Artificial Intelligence
A new algorithm called GROVER creates the most believable fake news in unlimited quantities.
- Trained on a large chunk of internet content, it creates news articles that are almost indistinguishable from those written by real people.
- While it might sound funny, it can also be worrying news. First of all, it may cost some journalists their jobs. Secondly, it can be used for online propaganda.
The founders claim that the best defence against Grover turns out to be... Grover itself, with 92% accuracy.
10. Google experiments with Doodle
Google is one of the leading companies in the field of AI. You can always go to their website and check out their latest experiments. For example, you can add your own doodles to the world’s largest doodle database to help their image recognition system improve.
These are the most striking cases of AI use for today. What shall we expect next? Driverless taxis? Robot-pets? We say anything is possible. Stay tuned with Magora’s blog to be aware of the latest trends in the AI world or be the next to lounge your own AI project.
Find out more facts about AI, ML and Deep Learning on the next pages of our blog: