An Approach to Joint Hybrid Decision Making between Humans and Artificial Intelligence

TL;DR

The paper tackles how to fuse human and AI decision-making in tasks with high stakes. It formalizes intelligence as decision-making competence balancing effectiveness and efficiency, applicable to both humans and AIs, and introduces a joint hybrid intelligence (si) framework with joint agent engineering (jae) at its core. A triad design space guides the integration of humans and AI, including demands analysis, task allocation, and joint-agent-pattern evaluation, exemplified by extended swarming in human–swarm interaction. The work provides a concrete, cross-disciplinary blueprint for designing, evaluating, and iterating intelligent joint systems across training, engineering, and interface design domains.

Abstract

Due to the progress in artificial intelligence, it is important to understand how capable artificial agents should be used when interacting with humans, since high level authority and responsibility often remain with the human agent. However, integrated frameworks are lacking that can account for heterogeneous agents and draw on different scientific fields, such as human-factors engineering and artificial intelligence. Therefore, joint hybrid intelligence is described as a framework abstracting humans and artificial intelligence as decision making agents. A general definition of intelligence is provided on the basis of decision making competence being applicable to agents of different sorts. This framework is used for proposing the interrelated design space of joint hybrid intelligence being aimed at integrating the heterogeneous capabilities of humans and artificial intelligence. At the core of this design space lies joint agent engineering with the goal of integrating the design subspaces operator training, artificial intelligence engineering, and interface design via developing joint agent patterns. The ''extended swarming'' approach to human-swarm interaction is discussed as an example of such a pattern.

Figures (11)

• Figure 1: An agent body with sensors, decision matrix, and effectors is embedded in a particular world.
• Figure 2: Intelligence depends on the observer constraints of agent purpose and agent-world situational contexts which constrain the required agent decision making competence. Adapted from Rockbach.
• Figure 3: Visualization of the situation space $\mathbb{S}$ as a collection of agent-world system variable states $s$ impacting the performance space $\mathbb{G}$ via the agent's decision making competence $c$. $G$ as a subset of the performance space describes the tolerated performance values of the system, while $g$ is a particular state in the tolerated goal space range. The agent's decision making competence $c$ determines how effective and efficient the agent is at finding goal states $g$ in the context of the situation space $\mathbb{S}$.
• Figure 4: High intelligence in terms of decision making competence requires a combination of effectiveness and efficiency.
• Figure 5: Blueprint of an intelligent architecture with fast reactive loops at the low level being modulated by slower but more capable cognitive loops. Adapted from Rockbach.

Theorems & Definitions (4)

• definition 1
• definition 2: Intelligence
• definition 3: Joint Hybrid Intelligence
• definition 4: Joint Agent Pattern