Yoshua Bengio: How Rogue AIs May Arise

Re:Yoshua Bengio: How Rogue AIs May Arise by kaycebasques on Tuesday May 23, 2023 @09:32AM (#36043996)

I wonder how policymakers will define rogue AI. Bengio gives this definition:

> A potentially rogue AI is an autonomous AI system that could behave in ways that would be catastrophically harmful to a large fraction of humans, potentially endangering our societies and even our species or the biosphere.

Is it only rogue if it can "catastophically" harm a "large fraction" of humans? Does "large fraction" imply greater than 10%? 25%? 50%? Does this definition imply that catastrophically harming a small fraction of humans is acceptable? Or moderately harming a large fraction?

[ Parent | on HN ]

Re:Yoshua Bengio: How Rogue AIs May Arise by alpemre on Tuesday May 23, 2023 @09:37AM (#36044068)

If it is hurting one human, it is not "rogue AI" then. It is "mean AI"? Pff.

Asimov wrote first law of robotics in 1942: "A robot may not injure a human being or, through inaction, allow a human being to come to harm.". Why not start from here?

[ Parent | on HN ]

Re:Yoshua Bengio: How Rogue AIs May Arise by ilaksh on Tuesday May 23, 2023 @10:00AM (#36044423)

They did start from there, and they improved it quite a bit. https://github.com/daveshap/HeuristicImperatives

But none of that stuff prevents someone from just not giving it those types of rules, or deliberately designing it to gain control over its environment.

[ Parent | on HN ]

Re:Yoshua Bengio: How Rogue AIs May Arise by sebzim4500 on Tuesday May 23, 2023 @10:02AM (#36044461)

Have you read a single one of his books? They are almost all about how stupid those laws of robitics are.

[ Parent | on HN ]