The Law of Stop

Verfassungsblog has published an essay titled 'The Law of Stop' arguing that AI governance needs a legal framework for stopping AI systems that pose serious risks or cause harm. The argument rests on an analysis of forty AI governance instruments. The essay's stated subject is the stop mechanism itself, which it argues should be set out in law.

The essay examines existing instruments, among them the EU AI Act. It notes that the act already requires providers of general-purpose AI models with systemic risk to report serious incidents to the AI Office. It also notes that those providers are not required to make such incidents public. The essay draws attention to the difference between the two: a serious incident can be reported to the AI Office without being disclosed publicly. The reporting duty, as the essay describes it, runs to a supervisory body rather than to the public.

The essay analyses forty AI governance instruments; the EU AI Act is the one the excerpt identifies by name. The excerpt does not name the other instruments or the jurisdictions they come from. The conclusion the essay states is that a legal framework for stopping AI systems that pose serious risks or cause harm is required. The essay presents the proposal as a requirement of AI governance, arrived at through the review of the forty instruments.