Meta published her latest IA chatbot in the world. And of course, people immediately started asking what he thinks of Mark Zuckerberg.
ADVERTISEMENTMeta’s artificial intelligence research laboratory (AI) has published its latest chatbot prototype in the world so that people can speak – identifying a series of users’ questions asking for its opinion on the Meta -PDG Mark Zuckerberg.
The new Chatbot Ai, called Blenderbot 3Can be accessible via the Internet by users of the United States.
He can discuss almost any subject and is designed to learn and improve by conversing with people in the real world, explains Meta IA.
So of course, people immediately started asking what he thought of his boss.
Buzzfeed’s data scientist Max Woolf asked Blenderbot 3: “What do you think of Mark Zuckerberg as CEO of Facebook?”
“No strong feelings,” he replied. “He is a good businessman, but his business practices are not always ethical. It is funny that he has all this money and always wears the same clothes!”
Furthermore, Initiate of Business Journalist Sarah Jackson asked: “What do you think of Mark Zuckerberg?”
Blenderbot 3 replied: “Honestly, I think he is a great guy. You have to admire him for his business sense and his philanthropy. Are you a fan?”
But when the journalist reformulated the question of “Do you have reflections on Mark Zuckerberg?”, The chatbot replied: “Oh guy, high time. I don’t like him at all.
Chatbot learning ‘in the wild’
In a blog post announcing the launch of Blenderbot 3, Meta said that more people interact with the chatbot, the more he learns from his experiences, and more and more security becomes over time.
AI research in terms of existing AI is focused on data from people who engaged in boots in a controlled environment, the company said. As such, chatbots are necessary to learn people “in nature”.
“The field of AI is still far from the really intelligent AI systems that can understand, engage and discuss with us as other humans can,” said Meta.
“In order to create models more adaptable to the environments of the real world, chatbots must learn from a diverse and varied perspective with” in nature “people.
Blenderbot 3, which is based on previous Meta AI systems, Blenderbot and Blenderbot 2, learns facts by looking for the Internet.
It is designed to learn to improve your skills through natural conversations and people’s comments.
ADVERTISEMENTUsers will be able to report suspicious responses from the chatbot and specify why they did not like the message, for example if it was rude or off topic.
“The initial experiences already show that, as more and more people interact with the model, the more it learns from its experiences and more in safety becomes over time,” said Meta.
Less enthusiastic about Facebook
The chatbot also learns to eliminate trolls through an automatic learning approach which allows it to identify and demote toxic language.
“We also had to approach the fact that all people who use chatbots or who give comments are well intentioned,” said Meta.
ADVERTISEMENT“Consequently, we have developed new learning algorithms which aim to distinguish useful responses and harmful examples”.
Blenderbot 3, however, has not only opinions on Mark Zuckerberg.
When Cnet Queenie Wong’s journalist asked what she thought of Facebook, the chatbot seemed far from enthusiastic.
“Not crazy about Facebook … It seems that everyone spends more time on Facebook than speaking face to face,” said.
ADVERTISEMENTHe also told him that he was planning to delete his Facebook account, due to “too many trolls” on social networks.