AI Reports · Harvard · 2026-10-19 · Blog / commentary
Is it already too late to reduce catastrophic AI risk?
A new Harvard Kennedy School Belfer Center paper, a 'pre-mortem', identifies current threats and vulnerabilities of frontier AI models and games out probable scenarios. It concludes it is not too late to act before the first major AI-enabled tragedy, though there is real cause for concern unless AI firms take decisive action. The authors argue the outcome will be shaped largely by the choices AI companies and governments make now.