Register for the event here

Alejandro Ortega will be speaking about a novel taxonomy and preparedness framework for dealing with loss-of-control situations. Despite increasing policy and research attention, existing LoC definitions vary significantly in scope and timeline, hindering effective LoC assessment and mitigation. To address this issue, Ortega and colleagues draw from an extensive literature review and propose a graded LoC taxonomy, based on the metrics of severity and persistence, that distinguishes between Deviation, Bounded LoC, and Strict LoC.

Ortega and colleagues model pathways toward a societal state of vulnerability in which sufficiently advanced AI systems have acquired or could acquire the means to cause Bounded or Strict LoC once a catalyst, either misalignment or pure malfunction, materializes. They argue that this state becomes increasingly likely over time, absent strategic intervention, and propose a strategy to avoid reaching a state of vulnerability. Rather than focusing solely on intervening on AI capabilities and propensities potentially relevant for LoC or on preventing potential catalysts, the authors introduce a complementary framework that emphasizes three extrinsic factors: Deployment context, Affordances, and Permissions (the DAP framework). Compared to work on intrinsic factors and catalysts, this framework has the unfair advantage of being actionable today. Finally, the authors put forward a plan to maintain preparedness and prevent the occurrence of LoC outcomes should a state of societal vulnerability be reached, focusing on governance measures (threat modeling, deployment policies, emergency response) and technical controls (pre-deployment testing, control measures, monitoring) that could maintain a condition of perennial suspension.

This talk will be taking place in the canteen at Meridian Cambridge (53-54 Sidney Street; open the front door and head all the way upstairs!).

All are welcome - please bring friends! We'll order in lunch.