Arvind Narayanan & Sayash Kapoor: AI as Normal Technology
Primary Source: AI Snake Oil / Princeton University Press
Authors: Arvind Narayanan & Sayash Kapoor · Year: 2024 · Publisher: Princeton University Press / Substack
Summary
A sharp demystification of AI hype and existential risk narratives. Narayanan and Kapoor argue that artificial intelligence should be conceptualized, analyzed, and governed as a “normal technology”—analogous to computing, databases, or electricity—subject to institutional inertia, empirical measurement, engineering trade-offs, and existing regulatory frameworks.
Key Points
- Rejection of Technological Exceptionalism: Demystifies the framing of AI as an uncontrollable metaphysical force or imminent existential threat, grounding it in mundane infrastructural realities.
- Predictive AI vs. Generative AI: Distinguishes between flawed predictive systems (e.g., recidivism scoring, hiring algorithms) and useful generative tools (e.g., coding assistants, writing aids).
- Institutional Governance: Argues that normal technological failures (bias, privacy violations, security flaws, environmental costs) can and should be addressed using established institutional, legal, and engineering oversight.