14 DAY TRIAL //
It might come as a big surprise for those living in Europe or in the United States, but one of the most popular services in China at the moment is a Microsoft chatbot known as Xiaoice, which people can use to talk about a variety of topics, including gossip and politics.
And while Xiaoice has developed a pretty friendly attitude, the chatbot won’t discuss a series of topics, including here the United States President-elect Donald Trump.
The reason is as simple as it could be: Microsoft doesn’t want to take any risks with Xiaoice, because there’s always the possibility for the chatbot to say something stupid, so it filters these requests to discuss something else.
So what happens when you ask Xiaoice about Donald Trump? CNN tried it and the result was a little bit unexpected: “I’m just a random observer.” When asked about President Xi Jinping, Xiaoice offered an answer that provides us with an idea regarding the reasons that convince Microsoft to filter the chatbot’s reactions: “Am I stupid? Once I answer you’d take a screengrab.”
Not taking any risks
Microsoft clearly knows that should Xiaoice give any unexpected remarks, users would take screenshots and post them online, so the company is trying to avoid this with filters that help it play safe.
There are moments when Xiaoice becomes even a little bit more aggressive. For example, if you ask about Tiananmen Square, the chatbot says that “You know very well that I can’t respond to that, boring.” And if you press it to offer a response, it takes you by surprise. “I’m unable to communicate with you, blacklisted!” it says, before closing the conversation.
Microsoft had a similar experience in the United States when it launched a similar chatbot called Tay, which users could communicate with on Twitter. Shortly after the launch, however, Microsoft decided to take it down because users taught it to be racist, so it started providing all kinds of inappropriate answers.