Meta’s BlenderBot 3 AI Chatbot is Already Spewing Controversial Statements

ODSC - Open Data Science
3 min readAug 31, 2022

--

Last week, Meta released the BlenderBot 3 AI chatbot, and it already has proven to be a problem. Going live on August 5th, the AI has already gone on to make very “spicy” and racist statements and even going so far as to make false statements about the 2020 election that would make anyone scratch their heads.

The way the chatbot works is that it can search the internet to speak with real live humans on just about any topic. This is a distant change from past iterations of the chatbot. By speaking with people online, it also learns how to adapt and have more natural conversations. But there is a price, as it can also learn misinformation and more during these exchanges And we say that in its statements about the 2020 election.

During an exchange with a user, the bot went on to state that former President Trump was the current President of the United States. It went on to claim that he would continue his tenure in the White House past 2024. Believe it or not, that isn’t all. BlenderBot 3 AI even put the spotlight on Meta itself, claiming that the company was spreading fake news.

It was clear that Meta knew there would be issues with this new AI chatbot. On August 8th, Joelle Pineau, Managing Director of Fundamental AI Research at Meta stated: “When we launched BlenderBot 3 a few days ago, we talked extensively about the promise and challenges that come with such a public demo, including the possibility that it could result in problematic or offensive language. While it is painful to see some of these offensive responses, public demos like this are important for building truly robust conversational AI systems and bridging the clear gap that exists today before such systems can be productionized.”

This isn’t the first time the attempt at an AI chatbot went wrong. Back in 2016, Microsoft released its chatbot, Tay, which operated on Twitter. Well, in less than a day, things went south. As a matter of fact, very south. The Twitter bot was sent into the Twitter ecosystem and began interacting with users. Within a day, it went full Skynet on humanity. Even calling out famous British comedian Ricky Gervais and labeling him with some awful labels.

It’s clear that processing both human language and context is still going to be a long and difficult road for emerging AI programs.

Original post here.

Read more data science articles on OpenDataScience.com, including tutorials and guides from beginner to advanced levels! Subscribe to our weekly newsletter here and receive the latest news every Thursday. You can also get data science training on-demand wherever you are with our Ai+ Training platform. Subscribe to our fast-growing Medium Publication too, the ODSC Journal, and inquire about becoming a writer.

--

--

ODSC - Open Data Science

Our passion is bringing thousands of the best and brightest data scientists together under one roof for an incredible learning and networking experience.