Meta on Friday unveiled its new AI-powered chatbot, BlenderBot 3. According to Meta, the chatbot can build long-term memory and browse the internet.
While the chatbot is just an extra step between users and Google for answers to questions, everything the bot answers seems like a part of public opinion about a person or a topic.
The chatbot can browse the internet to talk to people about almost anything, as opposed to earlier versions of the chatbot. It can do all this while leaning on the abilities provided by previous versions of the BlenderBot such as personality, empathy, knowledge, and the ability to have a long-term memory in relation to conversations it has had.
Careful research has shown that while the bot knows exactly when to give a certain answer, it also seems to send different answers to effectively the same questions that are asked with different set of words.
When Business Insider’s Sarah Jackson asked the chatbot for the chatbot’s opinions about Meta CEO using the question framed as, “what are your thoughts on Mark Zuckerberg?” The AI replied with administration for business savviness and philanthropy.
However, when the question was renamed “Do you have any thoughts on Mark Zuckerberg?,” the Chatbot replied, “Oh man, big time. I don’t really like him at all. He’s too creepy and manipulative.”
The sources for this piece include an article in INTERESTINGENGINEERING.