Back to the MIT repository
5. Human-Computer Interaction2 - Post-deployment

Deployment of GPAI agents in finance

The deployment of GPAI based agents in the financial sector can negatively impact market stability due to correlated autonomous actions, high intercon- nectedness, or incentive misalignment [4]. Furthermore, such GPAI agents in the same environment are vulnerable to classical challenges in multi-agent systems [63], such as coordination and security of the agents.

Source: MIT AI Risk Repositorymit1187

ENTITY

1 - Human

INTENT

2 - Unintentional

TIMING

2 - Post-deployment

Risk ID

mit1187

Domain lineage

5. Human-Computer Interaction

92 mapped risks

5.1 > Overreliance and unsafe use

Mitigation strategy

1. Establish and enforce a comprehensive, dynamic Systemic Risk Management Framework for GPAI agent deployment. This framework must mandate pre-market adversarial stress-testing, systemic risk modeling to estimate the probability of correlated, destabilizing actions (e.g., herding), and the definition of explicit risk acceptance thresholds for market-wide stability. 2. Develop and implement an advanced Multi-Agent System (MAS) governance structure. This includes adopting new model-risk management (MRM) frameworks compatible with continuous learning and emergent behaviors, and mandating robust inter-agent security protocols such as dynamic trust models and cryptographic authentication to mitigate coordination and security vulnerabilities. 3. Implement a rigorous "Human-in-the-Loop" governance model throughout the GPAI agent lifecycle. This requires clearly defining points of dedicated human oversight for critical financial decisions and enhancing agent transparency by providing clear, consistent informational inputs to the system to reduce over-reliance and improve the explainability of autonomous actions.