Former OpenAI board member Helen Toner said that AI regulators may only take sufficient action after a major real-world disaster [1].
The warning highlights a growing concern among governance experts that current policy efforts are insufficient to prevent catastrophic risks. If regulators continue to lag, the industry could face a crisis before meaningful safety guardrails are established.
Speaking in Washington, D.C., Toner said there is a disconnect between the perceived risks of artificial intelligence and the actual pace of government intervention [1]. She said that the AI governance community is increasingly concerned that a stark incident is required to shift the political will toward stricter oversight.
"A lot of people in [the AI governance space] worry that we need some kind of Chernobyl accident, some kind of real‑world disaster to wake up and to actually take action," Toner said [1].
Her comparison to the 1986 nuclear disaster suggests that regulatory frameworks often emerge as reactive measures rather than proactive protections. This pattern of "disaster-driven regulation" could be particularly dangerous given the scale and speed of AI deployment across global infrastructure.
Toner said that without such a wake-up call, the current trajectory of AI development may outpace the ability of policymakers to manage its risks [1, 2]. The lack of a concrete, catastrophic event may be creating a false sense of security among those tasked with creating laws to govern the technology.
“AI regulators may only take sufficient action after a major real-world disaster.”
Toner's assessment reflects a systemic tension in technology policy where the absence of a visible failure is often interpreted as a lack of risk. By invoking Chernobyl, she argues that the current regulatory inertia is not a sign of safety, but a lack of urgency that could lead to a catastrophic failure before the government implements necessary controls.
