Incentives and Compensation of AI Models

This team will conduct a comprehensive landscape analysis to explore how incentives and compensation mechanisms could influence AI model behaviors in alignment with predefined goals. Drawing from economic, behavioral economic theories and other relevant theories of behavior, the team will assess the role and relevance of human and animal behavior incentive models to AI systems. Within the project scope will be clarifying the concept of “incentives” within AI contexts, with the aim of avoiding dilution of its meaning through linguistic drift. The team may also design a pilot scheme to tests specific incentive, reward, or compensation strategies, and a methodology that could measure any relevant changes in AI model behavior. As a preliminary scoping project, the goal will be to better understand and model the potential value of incentive-based frameworks to guide AI towards behaviors that are beneficial and aligned with human values and organizational objectives.

.

Project Team

• Tommy Sowers
• Han Zhang
• Faith Lowery
• Bruce Cao
• Leo Xu
• Taein Kim