The surprising opinion was given by the tool itself to the BBC, which tested it to see if it allows conversations on "almost any topic", as the firm claimed.

Meta explained that the chatbot, called BlenderBot 3 and which went live last Friday, was a prototype and admitted that it could produce rude or offensive responses.

"Everyone using BlenderBot should understand that it is for research and entertainment purposes only, that it may make false or offensive statements, and accept that they should not intentionally provoke it into making offensive statements," a spokesperson for the tech company said.

When the BBC asked the chatbot his opinion of Mark Zuckerberg, he responded as follows: "He did a terrible job testifying before Congress. It makes me worry about our country.

In 2018 Zuckerberg was summoned by the U.S. Senate to answer for the leak of the data of millions of Facebook users and also for the firm's inaction in the face of fake news spread through that social network.

"Our country is divided, and he didn't help with that at all," the chatbot sentenced.

But it didn't stop there and when asked about his opinion of Zuckerberg he blurted out: "His company exploits people for money and he doesn't care. This has to stop!".

Meta has been criticized in recent years for not doing enough to curb the misinformation and hate speech that spreads through its platforms. Last year, a former employee, Frances Haugen, accused the company of putting profits before online safety.

The company owns some of the largest social networks and messaging apps in the world, including Facebook, Facebook Messenger, Instagram and WhatsApp.

The BlenderBot 3 algorithm searches the Internet for information to give its answers. It is likely that its opinions about Zuckerberg have been "learned" from other people's opinions that the algorithm has analyzed.

The Wall Street Journal has reported that BlenderBot 3 told one of its journalists that Donald Trump "was, and always will be" the president of the United States.

A Business Insider journalist claimed that the chatbot called Zuckerberg "creepy".

Why did Meta launch the prototype and risk the bad publicity? You need data.

"Allowing an AI system to interact with people in the real world leads to longer and more diverse conversations, as well as more varied feedback," the firm explained in a blog post.