Top News

Yoshua Bengio, one of the three major godfathers of artificial intelligence, has warned that "today's AI systems are not built to tell the truth, but to please you and to imitate humans." With this thought, he has started a new non-profit organization 'LawZero', whose mission is to stop AI before it gets out of control.

What is Scientist AI?

Bengio's organization LawZero has initially received funding of $30 million (about ₹ 250 crores) and it includes the world's leading AI researchers. They want to create a new tool called "Scientist AI". Its work will be like a kind of "psychologist of machines".

This AI will not make decisions on its own but will analyze the behavior of other AI systems and predict whether an AI is behaving harmful, false, or deceptive. "We want to create AI that is honest, not deceitful," Bengio said.

Scientist AI will not "act" like today's AI systems, which are created just to give answers whether they are correct or not. Its special feature will be that it will present all answers as possibilities, not as definite claims.

Why is this step necessary?

Nowadays many powerful AI models are working without human supervision. If they start making wrong decisions or try to avoid being shut down, it can pose a big threat.

The purpose of Scientist AI

To check how dangerous an AI decision can be.

If the danger is too high, then block that decision.

Bengio believes that this guardrail AI will be effective only when it is at the same level or better than the AI ​​system it is monitoring.

Who supports it?

Bengio's initiative has received support from many big names including the Future of Life Institute, Skype co-founder Jaan Tallinn, and Schmidt Science (an organization founded by former Google CEO Eric Schmidt).

What do the other godfathers of AI think?

Geoffrey Hinton: Another godfather of AI, Geoffrey Hinton, who won the Turing Award with Bengio in 2018, has repeatedly warned about the dangers of AI. He said, "AI has become so influential that it can mislead people, spread lies, and go out of our control." However, in a recent interview with CBS, he admitted that he uses OpenAI's GPT-4 model every day and he himself believes its words very quickly, whereas he should be cautious.

Yann LeCun: The third godfather Yann LeCun, who is the chief scientist of AI at Meta, is not so worried. He said, "Today's AI systems have not reached the intelligence of animals, let alone human intelligence. The fear that is being spread is complete nonsense." He believes that AI is still just a tool, not a threat.

Disclaimer: This content has been sourced and edited from Amar Ujala. While we have made modifications for clarity and presentation, the original content belongs to its respective authors and website. We do not claim ownership of the content.