A Framework for Understanding AI-Induced Field Change: How AI Technologies are Legitimized and Institutionalized

AI legitimacy varies dramatically by field: Different areas where AI is deployed—healthcare, facial recognition, finance, autonomous vehicles—face different levels of scrutiny and acceptance. This means there's no one-size-fits-all approach to responsible AI implementation; each sector needs tailored governance.

Weak institutional rules create a double-edged sword: Fields with minimal existing regulations and standards can adopt AI quickly, but this speed comes at a cost—these areas are more likely to face public backlash and contestation. The lack of clear guardrails makes rapid deployment riskier.

AI reshapes institutions while institutions constrain AI: The relationship is bidirectional—AI systems change how organizations operate, but existing rules, standards, and regulations simultaneously determine how much change can actually happen. This creates tension between innovation and stability.

Underdeveloped governance infrastructure is a major bottleneck: The paper identifies that most AI-implementation fields lack adequate institutional infrastructure (standards, rules, regulations), which could become a significant obstacle as AI becomes more widespread. This suggests urgent need for proactive governance development.

Institutional legitimacy must precede mass adoption: For AI to be broadly institutionalized across society, the underlying governance frameworks need to be established first—not after problems emerge. This emphasizes the importance of getting governance right early.