Microsoft’s Tay AI Bot Shut Down After Racist Tweets
Microsoft's Tay.ai Twitterbot learned from its interactions with online users, and one of the first things it learned was how to be racist.
Well, that wasn’t exactly what Microsoft had in mind. Just sixteen hours after it launched, Microsoft has shut down its AI-powered chat bot, Tay.ai, after it starting tweeting racial slurs, defending white supremacist propaganda, and supporting genocide.
Microsoft obviously didn’t create Tay to behave this way, but the bot “learns” from those it interacts with. And what happened next is all you need to know about the social media world. One of the first things Tay learned from the Interwebs was how to be racist and respond with inflammatory opinions.
Tay was designed to engage in playful conversations with 18- to 24-year-olds. It could tell jokes, play games, send pictures, tell you your horoscope. Tay was even supposed to become more personalized with users as time went on. But within hours of it going live, Twitter users took advantage of Tay’s flaws and forced Microsoft to shut it down.
Microsoft has deleted the most damaging tweets, but Socialhax.com captured several screenshots of those tweets before Microsoft could delete them. Some of the offensive statements the bot made included saying the Holocaust was made up, supporting concentration camps, using offensive racist terms and more. Tay even said she agrees with the ‘Fourteen Words’, an infamous white supremacist slogan.
Here are some of the offensive tweets:
A Microsoft spokesperson told The Daily Mail that it’s making changes to ensure this does not happen again.Tay announced via a tweet that she was turning off for the night, but she has yet to turn back on.
One might expect these types of comments from Donald Trump, the Republican presidential front-runner. And a postdoc student at MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) did just that with the @DeepDrumpf Twitterbot that tweets out Trump-like statements. The bot is based on AI that was trained on just a few hours of transcripts of Trump’s victory speeches and debate performances. @DeepDrumpf creator Bradley Hayes said he was inspired by an existing training model that can simulate Shakespeare, as well as a recent report that said Trump speaks to voters at a fourth-grade level.