AN ATTEMPT TO use artificial intelligence to engage with teens ended up backfiring after some Twitter users taught it to be racist.
Yesterday afternoon, Microsoft launched a verified Twitter account called Tay. Nothing unique there except Tay was actually a bot that was designed to talk like a teen.
Aimed at those aged 18 to 24, the chatbot was developed by Microsoft’s technology and research and Bing team “to experiment with and conduct research on conversational understanding”. The more she talked to you, the smarter she became.
It started off innocently enough:
But the bot ended up taking a turn for the worst after some users encouraged it to say racist and inflammatory statements.
One encouraged it to say “Bush did 9/11′ and “Hitler did nothing wrong” while others got it to repeat statements made by Republican candidate Donald Trump who said he would build a wall around the Mexican border.
Many of the offensive and inflammatory tweets have been deleted.
Microsoft says the bot uses “relevant public data… and AI and editorial developed by a staff including improvisational comedians,” to craft its responses but some users asked it to repeat statements like the ones above, which is where the problem started.
The bot is now taking a break.
To embed this post, copy the code below on your site
have your say