Some users of this chatbot have decided to do experiments to measure the responsiveness of this artificial intelligence
the most advanced AI in Meta who has confessed that he doesn’t like his ‘boss’, Mark Zuckerberg , at all , calling him a ” manipulator ” and “unethical” man. .
A few years ago, movies of robots attacking the human race, such as ‘Terminator‘ or ‘I, Robot’, were considered ‘science fiction’.
However, the evolution of technology has made possible the development of AI (artificial intelligence) devices capable of responding to human concerns with a strange and disturbing naturalness.
Now, overcoming Siri‘s bad jokes, the BlenderBot 3 has appeared.
According to the information published by ‘Business Insider’, the BlenderBot 3 chatbot has referred to the CEO of Meta as ” a bad person ” and has also indicated that he does not ” like him very much “, in response to questions asked by users about Mark Zuckerberg.
However, just like with Siri or Alexa , these answers can vary depending on the context and each question session.
Some users of this chatbot have decided to carry out experiments to measure the responsiveness of this AI, including Max Woolf , a scientist at ‘BuzzFeed‘, and the journalist for ‘Business Insider’, Sarah Jackson .
Max Woolf decided to ask the BlenderBot 3 the following: “how do you feel about Mark Zuckerberg as CEO of Facebook?”, the chatbot replied: “ No strong feelings.Mark Zuckerberg is a businessman , but his business practices are not always ethical. It’s funny that he has all this money and still wears the same clothes!”, causing surprise by this unexpected response.
On the other hand, journalist Sarah Jackson got an equally if not more surprising response from this chatbot when she asked, “Do you have any thoughts on Mark Zuckerberg?” to which the AI replied, “Oh man, great time. I really don’t like it at all. It’s too creepy and manipulative .”
However, other people who asked him the same questions received much more positive responses about Zuckerberg, so this type of negative assessment can be totally arbitrary or related to users’ internet search parameters , so the algorithm would change and so would their responses.
They accuse META AI of being racist and promoting conspiracy theories
In addition to shocking his ‘personal appreciation’ in relation to Mark Zuckerberg, it has also been discovered that the BlenderBot 3 had been ‘contaminated’ with harmful and negative ideas after interacting with real humans.
Among these ideas and responses, we can highlight some racist and xenophobic comments that have been influenced by groups that promote this language of hate on social networks, among which anti- Semitism stands out .
In addition, they also emphasize that this chatbot defends the theory that Donald Trump is still the president of the United States and not Joe Biden.
Faced with this situation, the directors of Meta have warned people who are testing
BlenderBot 3 about this, and have explained that they are working to develop additional protections so that the AI is not contaminated with the ideas imposed by its users, as happened previously with a Facebook AI that labeled several black people as ‘primates’ .