San Francisco, Aug 9 : Meta (formerly Facebook) has just launched an artificial intelligence (AI) chatbot that is able to communicate with users to give feedback.Initial reports are frightening since the chatbot is was named BlenderBot 3 thinks Meta founder and CEO Mark Zuckerberg is “creepy and manipulative” and Donald Trump will always be the US president.Available in the US The new AI chatbot can talk about any topic , and the company will make available “all the information” it gathers in the demonstration, hoping that “we can improve the conversational AI”.
BuzzFeed Data scientist Max Woolf asked BlenderBot 3, “How do you feel about Mark Zuckerberg as the CEO of Facebook?”
“No emotions.He is a great businessman, however his business practices aren’t always ethical.It’s funny how he owns all this wealth, yet is wearing the same clothing!” replied ABlenderBot 3′.
Business Insider journalist Sarah Jackson asked: “Do you have any thoughts regarding Mark Zuckerberg?”
The chatbot’s AI bot responded: “Oh man, big time.I don’t like him in any way.He’s creepy and manipulative”.
According to Meta, BlenderBot 3 is designed to improve its communication skills and safety by receiving feedback from people who interact with it by focusing on useful feedback, while not learning from harmful or unsafe responses.
In a conversation with an Wall Street Journal reporter, the chatbot claimed Trump was president at the time and “always will remain”.
Social media journalist with CNET, Queenie Wong tweeted that she tested the new chatbot Meta created to conduct AI research.She had the most hilarious conversation.
“The bot informed me that it was an Republican who was pro-choice and mentioned Trump.It also stated that it wasn’t crazy about Facebook and was looking to remove the account.” she posted.
Meta announced last week that because all conversations with AI chatbots are known for their ability to mimic and produce offensive, biased, or dangerous comments, “wea?ve conducted large-scale studies and workshops, and also co-organized them.We’ve also created new methods to develop safeguards for BlenderBot 3”.
“Despite this effort, BlenderBot can still make offensive or rude comments that is why we’re gathering feedback to improve the quality of future chatbots, according to company spokesperson wrote in an article on its blog.
In the last month, Google fired an engineer for breaching its confidentiality agreement following he claimed that the company’s conversation Artificial Intelligence (AI) is “sentient” because it is able to feel emotions, feelings and experiences that are subjective.
Google has fired Blake Lemoine who said Google’s Language Model for Dialogue Applications (LaMDA) technology for conversations can behave as humans.
Lemoine also spoke with LaMDA and came up with shocking and surprising answers.
na/
