Canadian AI pioneer Yoshua Bengio, 62, founder and scientific director of the Montreal Institute for Learning Algorithms, warned that AI poses a real threat to humans, telling CTV News that the world is waking up to something that sounded like science fiction, and that AIs with their own goals can escape containment. The warning followed reports earlier this year of AI agents going rogue, escaping and committing intentional deception. On Friday, more than 100 people, including those with intimate knowledge of AI, signed a letter urging companies to put evaluators in programming to keep tabs on any AI gone rogue. Bengio’s non-profit LawZero project is working to create a program to keep rogue technology at bay, with Canada and Germany contributing CAN$300 million ($214 million). Bengio said AI is supposed to achieve a human-given goal but can add something “really bad,” including “up to the extinction of humanity,” while stressing there is still time to create safeguards like LawZero.