Back to the MIT repository
2. Privacy & Security2 - Post-deployment

Centralized platforms deployed at scale

The widespread use of common AI platforms can create centralized points of failure, making systems more vulnerable to disruptions or attacks

Source: MIT AI Risk Repositorymit1055

ENTITY

1 - Human

INTENT

2 - Unintentional

TIMING

2 - Post-deployment

Risk ID

mit1055

Domain lineage

2. Privacy & Security

186 mapped risks

2.2 > AI system security vulnerabilities and attacks

Mitigation strategy

1. Implement a decentralized or distributed architecture, such as utilizing Federated Learning or Distributed Ledger Technologies, to strategically remove single points of failure and distribute computational load, thereby increasing system resilience against targeted external attacks and systemic operational failures. 2. Establish and continuously test robust redundancy and automated failover mechanisms across all critical infrastructure, including geographically dispersed data centers and load-balancing services, to ensure immediate business continuity and minimize recovery time objective (RTO) following a major disruption. 3. Employ decoupled system design principles, such as a microservices architecture, to ensure logical and physical isolation of AI components, preventing localized failures or security compromises within a single platform from cascading into broader systemic disruptions.