Pedro Luis Graterol

I know talking about the AI threat as extinction-level is a hot topic this week. However, an even more likely scenario that we should also be worried about is an AI-caused 9/11 style event - something that causes mass destruction, kills thousands or tens of thousands of people, and reshapes global politics for a generation. Various cyber threats to critical infrastructure can absolutely do that level of damage. My p(doom) - the probability of AI causing human extinction - is less than 1%. Still really scary, but far from the most likely scenario. My p(destruction) of a catastrophic 9/11-level event caused by AI at some point in the next decade hovers right around 40-50% (some of that depends on how you specify the vague definition of “9/11-level”). The doomsday stuff drives the debate this week, but it’s a good day to remember that events don’t have to be extinction-level to be absolutely devastating and world-changing. Technology companies, politicians, regulators, and citizens should also work to prevent or mitigate the AI threats that most of us survive but regret not working harder to stop in hindsight.