PALO ALTO, California — Artificial intelligence pioneer Yoshua Bengio plans to develop a monitoring AI that prevents other deep-learning programs from going rogue and resorting to behaviors harmful to humans to achieve their set goals.
Alarmed by the rapid advance of AI capabilities and the lack of global regulation, Bengio is setting up a nonprofit organization dedicated to researching AI safety. He aims to promote an “ethical AI” that can predict and prevent harmful behaviors.
