Public Interest AI
Public interest AI is the governance alternative discussed in Kate Crawford: Mapping Empires during Kevin Kelly’s Q&A with Kate Crawford. The source asks whether AI could be publicly accessible, publicly funded, publicly accountable, and technologically sovereign without simply reproducing the resource-intensive private AI stack.
Crawford’s answer is conditional rather than purely enthusiastic. Public AI must be judged through AI Metabolic Infrastructure: renewable energy, consent-based data, democratic oversight, and narrow domain fit matter as much as access. A public system that merely duplicates private scaling races could worsen energy, water, mineral, and data extraction.
Key Claims
- Public access is not enough if the system repeats the same extractive data and infrastructure model.
- Consent-based data frameworks are part of public legitimacy.
- Renewable energy and sustainability requirements are core design constraints, not public-relations additions.
- Public AI should focus on domains where AI has demonstrated value, such as translation, astronomy, climate research, and data-gap analysis.
- Democratic oversight and AI Literacy Against Worship are needed so citizens can understand and shape AI rather than only consume it.
Connections
- Kate Crawford and Kevin Kelly - source exchange that defines the idea.
- AI Metabolic Infrastructure - resource constraint that public AI must address.
- AI Literacy Against Worship - public understanding needed for democratic oversight.
- AI Alignment Governance and AI Governance And Compliance - governance branches adjacent to public accountability.
- Human Flourishing Profit - neighboring Long Now idea that technology should be judged by human flourishing, not only accounting returns.
- Public Service Journalism - adjacent public-interest institution branch around civic information.