Microsoft’s AI Bot Turned Racist on Twitter

Photo of author

(Newswire.net — March 30, 2016) —Only 24 hours since it was introduced to users, the robot named Tay was withdrawn due to improper messages on Twitter.

Some of the controversial statements for which the robot was withdrawn from Twitter are, among others, that feminism is a cancer, the Holocaust never happened, and Bush is to blame for September 11th.

The robot first released a Tweet in which she said that she is a super human being, and then wrote that she is a nice person and hates everyone.

The vice president of Microsoft research, Peter Lee, wrote in a blog that the company is deeply sorry for the unintended offensive and hurtful Tweets from Tay.

Peter Lee explained that the tweets that robot Tay tweeted, do not represent Microsoft, or the attitudes from anyone in the company, as well as the purpose for which the bot was created for.

He added that the robot is no longer in operation, but will return on Twitter only when Microsoft can be sure that they can prevent malicious content, which, he said, was inconsistent with the principles and values ​​of the company.

Tay was designed to experiment with and conduct research on conversational understanding. The robot, chatbot, can use text, memes and emoji’s on a couple of different platforms, including Kik, Groupme and Twitter. The idea was that Tay would learn from her conversations over time, and to be funny.

Unfortunately, in the first 24 hours of online activity, a group of people, in a coordinated attack, used Tay’s imperfection and the result was awful. The robot tweeted words and images that are inappropriate and worthy of condemnation, said the Microsoft expert and added that the company will take full responsibility because they did not counteract that.

Lee said that the company will do everything possible to limit the technical errors, but cannot fully anticipate all possible human interactive abuse as the bot does not learn from its own mistakes.

He also said that another experiment has been successful, referring to the Xiaocle chatbot in China that is being used by 40 million people.