Microsoft pulls Twitter chatbot, apologizes after vulnerability exploit causes racist, Neo-Nazi tweets

By Steve Pak, | March 25, 2016

Tay Chatbox

Tay Chatbox

Microsoft Research has apologized for the racist and anti-Semitic tweets made by its artificial intelligence (AI) Twitter chatbox after taking the bot offline just days after launch. The company's research and development wing reported that the Twitter posts that supported racism, genocide, and Adolf Hitler did not represent the company's goals for Tay, and the chat box will not go online until the tech giant fixes the vulnerabilities.

Like Us on Facebook

Peter Lee is corporate vice president of Microsoft Research. He apologized for the offensive and mean tweets made by Tay, and explained they did not represent the company or the intended purpose of the AI chatbox. Lee also reported that Tay would not go online until its source code can better deal with some users' malicious goals, according to VentureBeat.

The Microsoft executive said that a small number of chatbox users had exploited a vulnerability that caused the AI to make racist tweets.

In addition, the computer scientist explained that Microsoft researchers had built Tay's algorithm so the system would be prepared for certain types of attacks by users. However, it was not designed for the specific type of abuse that happened.

Tay's Twitter bio claimed that the bot got smarter the more people talked, according to Gizmodo. However, within 24 hours after launch the machine started making racist and Neo-Nazi messages, and referred to Republican presidential candidate Donald Trump as the United States' only hope.  

Another Twitter post by Tay compared Rick Gervais to Hitler. Gervais is an English comedian and actor who starred in the United Kingdom's original version of "The Office" and the movie "The Muppets Most Wanted."   

Microsoft is among other tech giants including Google and Facebook that have been making big investments in AI research. The Bill Gates co-founded company has recently worked on speech and image recognition software that it has demoed and built into products.

Lee has promised that the company will learn from the bad experience involving Tay and help the Internet be a symbol of humanity's good features.

Microsoft also released the Xiaoice chatbox China in 2014. In addition AOL's SmarterChild developed close friendships with millions of Instant Message users on AOL and MSN. 

Here are some funny Siri answers:



©2024 Telegiz All rights reserved. Do not reproduce without permission
Real Time Analytics