Framework for Building AI Agent-Human Relationships

The AOSM (Agent Oriented Systems Modeling) framework provides a structured approach to designing human-AI teaming relationships, ensuring effective collaboration, interdependence, and trust. The framework consists of the following key components:

1. System Structure for Human-AI Teaming

• A human-AI team is an interdependent system where AI and humans share goals.
• The system consists of: Human team members (users/operators) AI agents (intelligent systems) A user interface (for interaction) A machine component (optional, such as an automated vehicle) An external environment (context of operation)

📌 Example: In adaptive cruise control, the AI agent adjusts speed based on environmental inputs, while the human steers — demonstrating interdependence.

2. Principles for Effective Human-AI Teaming

The document defines four key principles:

(a) Interdependence

• AI and humans must rely on each other’s strengths while compensating for weaknesses.
• AI must support human tasks, not replace them.

📌 Example:

• AI-assisted medical imaging enhances human diagnosis by flagging potential anomalies, while doctors provide the final judgment.

(b) Observability, Predictability, and…