AI collaboration dynamics

Radboud 2023-2024

This project explores the integration of Question-Answering Systems (QAS), such as GPT-4, into fundamental physics research, focusing on high-energy physics and astronomy. It aims to harness AI-powered interfaces to augment human cognitive capabilities, facilitating innovative interactions with complex data. Emphasizing the collaborative nature of human-AI teams, the project departs from the autonomous systems approach to explore how scientific discovery is intertwined with collaboration dynamics and epistemic authority distribution in these teams. It investigates the potential of QAS in enabling new physics concepts discovery, through automatic idea generation and identifying new research avenues. By examining collaboration dynamics, defining roles, establishing communication channels, and ensuring the interpretability of AI systems, the project seeks to optimize the integration of QAS into research teams. Additionally, it addresses philosophical questions regarding the evolving nature of scientific discovery and the delegation of epistemic authority to QAS. The goal is to lay the groundwork for responsibly integrating AI into scientific research to synergize with human intellect and drive scientific advancements.