Researchers at the University of Toronto have amassed hundreds of signatures on an open letter calling for improved AI safety protocols and oversight, warning that recent high-profile security incidents could foreshadow far more damaging failures without intervention.
