The Role of AI in Facilitating Interdisciplinary Collaboration: Evidence from AlphaFold

The acceleration of artificial intelligence (AI) in science is recognized and many scholars have begun to explore its role in interdisciplinary collaboration. However, the mechanisms and extent of this impact are still unclear. This study, using AlphaFold’s impact on structural biologists, examines how AI technologies influence interdisciplinary collaborative patterns. By analyzing 1,247 AlphaFold-related papers and 7,700 authors from Scopus, we employ bibliometric analysis and causal inference to compare interdisciplinary collaboration between AlphaFold adopters and non-adopters. Contrary to the widespread belief that AI facilitates interdisciplinary collaboration, our findings show that AlphaFold increased structural biology-computer science collaborations by just 0.48%, with no measurable effect on other disciplines. Specifically, AI creates interdisciplinary collaboration demands with specific disciplines due to its technical characteristics, but this demand is weakened by technological democratization and other factors. These findings demonstrate that artificial intelligence (AI) alone has limited efficacy in bridging disciplinary divides or fostering meaningful interdisciplinary collaboration.

💡 Research Summary

The paper investigates whether artificial intelligence (AI) can meaningfully bridge disciplinary divides by focusing on AlphaFold, the deep‑learning system that revolutionized protein‑structure prediction. Using the Scopus database, the authors extracted 1,247 AlphaFold‑related articles and 7,700 unique authors. Each paper was classified as an “AI adopter” (i.e., the study explicitly used AlphaFold) or a “non‑adopter” (traditional methods). Authors’ institutional affiliations were mapped to OECD research fields, allowing the authors to identify interdisciplinary collaborations whenever co‑authors came from different fields.

To isolate the causal impact of AI adoption, the study combined propensity‑score matching (PSM) with a difference‑in‑differences (DiD) framework. PSM balanced observable pre‑treatment characteristics such as institutional size, funding levels, and author seniority between adopter and non‑adopter groups. The DiD model then compared changes in interdisciplinary co‑authorship rates before and after AlphaFold’s release.

The results reveal a modest increase of only 0.48 percentage points in collaborations between structural biology and computer science after AlphaFold adoption. No statistically significant changes were observed for collaborations involving other disciplines (e.g., physics, chemistry, medicine). The authors interpret this pattern as evidence of a “technology‑demand matching” effect: AI tools tend to generate interdisciplinary demand only with fields that share the tool’s technical requirements. Moreover, the rapid democratization of AlphaFold—its open‑source release and cloud‑based accessibility—appears to have diluted any initial surge in cross‑disciplinary activity.

The paper concludes that AI alone has limited power to dissolve entrenched disciplinary boundaries. While AI can create new points of contact between specific fields, broader interdisciplinary integration requires complementary organizational and policy measures, such as targeted funding incentives, joint training programs, and shared research infrastructure. The authors acknowledge several limitations: reliance on Scopus field classifications, measurement of collaboration solely at the publication level, and binary treatment of AI usage. They suggest future work should employ richer collaboration metrics (e.g., grant co‑funding, patent co‑ownership), examine other AI platforms, and explore longitudinal dynamics at the project stage. Overall, the study provides a nuanced, data‑driven counterpoint to the prevailing narrative that AI is a universal catalyst for interdisciplinary science.