ARTFEED — Contemporary Art Intelligence

HAAS Framework for Adaptive Human-AI Task Allocation

ai-technology · 2026-05-06

A new framework called Human-AI Adaptive Symbiosis (HAAS) addresses the challenge of distributing work between humans and AI systems in organizational design. Unlike binary approaches, HAAS allows for shared and complementary roles based on context, fatigue, and stakes. The framework combines a rule-based expert system for governance constraints and a contextual-bandit learner that selects collaboration modes from outcome feedback. Task-agent fit is evaluated through five cognitive dimensions and a five-mode autonomy spectrum, ranging from human-only to fully autonomous. The framework is implemented for software engineering and manufacturing.

Key facts

  • HAAS stands for Human-AI Adaptive Symbiosis
  • Framework combines rule-based expert system and contextual-bandit learner
  • Five cognitive dimensions used for task-agent fit
  • Five-mode autonomy spectrum from human-only to fully autonomous
  • Implemented for software engineering and manufacturing
  • Addresses binary choice limitation in task allocation
  • Governance constraints enforced before learning
  • Collaboration modes selected from outcome feedback

Entities

Sources