How large language models can reshape collective intelligence

Area of Science:

• Social Sciences
• Computer Science
• Information Science

Background:

• Collective intelligence enables groups to outperform individuals through distributed cognition and coordination.
• Information technology has historically supported collective intelligence via prediction markets, forums, and crowdsourcing platforms.
• Large language models (LLMs) represent a significant technological shift in online information processing.

Purpose of the Study:

• To examine the opportunities and challenges presented by large language models for collective intelligence.
• To synthesize interdisciplinary perspectives on the impact of LLMs on group problem-solving.
• To identify potential benefits, risks, and policy considerations related to LLMs and collective intelligence.

Main Methods:

• Interdisciplinary review synthesizing perspectives from industry and academia.
• Analysis of how LLMs transform information aggregation, access, and transmission.
• Identification of research questions and policy-relevant considerations.

Main Results:

• LLMs present unique opportunities and challenges for collective intelligence.
• The transformation of online information dynamics by LLMs requires careful consideration.
• Potential benefits and risks associated with LLMs in collective settings are highlighted.

Conclusions:

• A closer examination of how LLMs affect humans' ability to collectively tackle complex problems is crucial.
• Understanding the interplay between LLMs and collective intelligence is essential for future societal and organizational success.
• Further research is needed to navigate the evolving landscape of AI-enhanced collective intelligence.