Microsoft again testing the AI chat bot water

Microsoft is once again testing the AI chat bot waters with Zo, a seemingly reworked bot that was recently discovered on Kik by Twitter

user Tom Hounsell.

As you may remember, Microsoft’s first at bat – a chat bot named Tay – didn’t even last a full 24 hours before the Redmond-based company took it offline. Turns out, letting an impressionable chat bot designed to mimic a 19-year-old girl loose across multiple platforms on the Internet isn’t all that great of an idea.

Within hours, Tay was spewing racist comments left and right, offending hordes of people in the process.

According to MSPoweruser, Zo is best described as a censored version of Tay or an English variant of Xiaoice, Microsoft’s Chinese chat bot. It’s apparently pretty good at normal conversation although if you try to steer the conversation toward a controversial topic like politics, it’ll politely decline to engage.

Limiting Zo to Kik is likely a wise move as well. This gives Microsoft an opportunity to further fine-tune the bot before opening it up to popular social networks with more active users like Twitter and GroupMe. How long that’ll take, however, is anyone’s guess, especially considering Microsoft hasn't yet officially announced Zo (someone just stumbled upon it).