The Missing Colleague: AI and Informal Intellectual Collaboration

A scientific paper has two teams: its named authors and the invisible network of colleagues
who provide feedback, criticism, and advice. Generative AI can increasingly provide similar
intellectual assistance without requiring human interaction. Does AI therefore make the invisible
team smaller? Combining AI-text classifications with acknowledgments from 14,164 NBER
and arXiv economics working-paper records, we find that AI-flagged manuscripts acknowledge
approximately 21% fewer informal collaborators in specifications with researcher fixed effects
and extensive controls - equivalent to roughly one fewer intellectual contributor per paper. The
negative association grows in magnitude with AI-intensity. We develop a theoretical framework
identifying two channels through which AI may substitute for human intellectual input through
two channels. First, an expertise-substitution channel: AI may replace domain-specific feedback
traditionally obtained from colleagues working on similar research topics. Second, an access-cost
channel: AI may reduce the need to seek feedback from colleagues who are more costly to
reach. We find suggestive evidence for both channels. Consistent with expertise substitution,
AI-flagged manuscripts acknowledge fewer same-topic contributors. Consistent with reduced costs
of accessing expertise, they acknowledge fewer first-observed contributors and fewer contributors
from other institutions. Our findings suggest a potentially important organizational trade-off: AI
may substitute for the intellectual contributions of colleagues without reproducing the professional
relationships through which knowledge is exchanged, ideas circulate, and future collaborations
emerge.