Moneycontrol PRO
you are here: HomeNewsTrends

What Meta's new AI bot said about CEO Mark Zuckerberg in chats with humans

Meta's BlenderBot 3 is a publicly available chatbot that can talk about virtually anything.

August 06, 2022 / 04:03 PM IST
Mark Zuckerberg's Meta is aiming to research in conversational artificial intelligence.

Mark Zuckerberg's Meta is aiming to research in conversational artificial intelligence.


"Mark Zuckerberg's practices are not always ethical". "He can be creepy and controlling".  These are not the words of a disgruntled Meta employee but of an artificial intelligence-powered chatbot recently unveiled by the company.

BlenderBot 3 is a publicly available chatbot that can talk about virtually anything. "Capable of searching the internet, it’s designed to learn how to improve its skills and safety through natural conversation," Facebook's parent company said in a blog post on Friday.

People can chat with the bot on a website, though that option is available only in the United States for now.

Such attempts led to some interesting but inconsistent responses. Max Woolf, a Buzzfeed data scientist, asked the bot what it thought about Meta Chief Executive Officer Mark Zuckerberg.

 

"No strong feelings," the bot responded. "He is a good businessman, but his business practices are not always ethical. It is funny that he has all this money and still wears the same clothes!"

In a chat with another user, the bot had a harsher opinion of Zuckerberg.

To a third person, the bot said: "He (Zuckerberg) seems to care more about his image than the company itself."

In a fourth chat, the bot's words changed drastically. It described Zuckerberg as one of the best people on Earth.

 

 

Meta has acknowledged that its bot may make contradictory or inappropriate statements. The company said it hopes to improve its models using feedback.

Meta hopes to promote conversational AI research to eventually build computers that people can engage with in "helpful and interesting ways".

But as ethical and safety questions surrounding AI remain, Meta advised those engaging with its bot not to disclose personal information.

"If a user accidentally includes personal information in their chat, they can decide not to opt-in to storing the data for research at the end of a session," the company added. At that point the conversational data will be permanently deleted."
Moneycontrol News
first published: Aug 6, 2022 03:57 pm
Sections
ISO 27001 - BSI Assurance Mark