Geoffrey Irving, co-founder and chief scientist at the AI nonprofit Resolution, told Reuters that “the risk is ramping up pretty fast,” arguing that AI companies are “overplaying the extent to which this is a pure coordination problem” and “could just stop unilaterally” if they judged a specific development too dangerous. DeepMind's Neel Nanda separately put the odds of AI-driven human extinction at 10% or higher, a figure he called “ridiculously high” in its own right.
Both researchers work inside or closely alongside the labs they're implicitly criticizing, which is precisely what gives their comments more weight than similar warnings from outside academics: they're not speculating about what frontier labs might be doing, they're describing what they've observed from direct proximity to it.