System Resilience
A concept originating from engineering and public administration, referring to a system's ability to maintain core functions, recover quickly, and learn from failures, attacks, or disruptions. The core assumption is that incidents are inevitable, and the key question is whether the system can withstand shocks without collapsing. In the AI context, it focuses on whether the infrastructure and public services on which AI deployment depends possess buffering, isolation, and degraded-mode operation capabilities.
Timeline (last 90 days)
Cathy Wu's research emphasizes system-level resilience, focusing on how the entire transportation network redistributes traffic and maintains service levels when nodes fail.
The UK government is elevating AI safety governance from a single model-level alignment to the height of national resilience (system resilience).
The regulatory logic of system resilience governance shifts from focusing on 'whether the model is safe' to focusing on 'whether the entire system is resilient under scaled AI deployment'.
System resilience is a concept originating from engineering and public administration, referring to a system's ability to maintain core functions, recover quickly and learn when it encounters failures, attacks or disruptions.
Under the system resilience paradigm, AI risk responsibility is distributed among deployers, infrastructure operators and regulators, and governance tools become stress testing, redundant design, isolation mechanisms and incident response plans.
The governance logic of system resilience is closer to the macroprudential regulation of the financial system than to mere product compliance.
All Facts (6)
Cathy Wu's research emphasizes system-level resilience, focusing on how the entire transportation network redistributes traffic and maintains service levels when nodes fail.
50%UnverifiedThe UK government is elevating AI safety governance from a single model-level alignment to the height of national resilience (system resilience).
50%UnverifiedThe regulatory logic of system resilience governance shifts from focusing on 'whether the model is safe' to focusing on 'whether the entire system is resilient under scaled AI deployment'.
50%UnverifiedSystem resilience is a concept originating from engineering and public administration, referring to a system's ability to maintain core functions, recover quickly and learn when it encounters failures, attacks or disruptions.
50%UnverifiedUnder the system resilience paradigm, AI risk responsibility is distributed among deployers, infrastructure operators and regulators, and governance tools become stress testing, redundant design, isolation mechanisms and incident response plans.
50%UnverifiedThe governance logic of system resilience is closer to the macroprudential regulation of the financial system than to mere product compliance.
50%