CONCEPT

Clogger

The 2023 thought experiment by Archon Fung and Lessig — an AI system designed to maximize electoral victory through personalized, adaptive persuasion at scale, hypothetical in form but technically trivial to build.

Fung and Lessig's 2023 essay in The Conversation introduced 'Clogger' as a hypothetical AI campaign system that would combine the capabilities already demonstrated by large language models, micro-targeting infrastructure, and engagement optimization to produce personalized persuasion at unprecedented scale and sophistication. The point of the thought experiment was that every component already existed. The system was not speculative; it was an obvious near-term application of existing technology. The essay argued that the deployment of such systems was not hypothetical and that democratic societies needed to decide, before the systems were fully operational, what governance framework would apply. Clogger has since become a reference point in AI policy debates as the canonical illustration of how AI can be weaponized against democratic deliberation without any novel technical capability.

In The You On AI Encyclopedia

The argument operates at the intersection of Lessig's two longest-running concerns: institutional corruption in democratic politics (the subject of Republic, Lost) and the governance of digital architecture (the subject of Code). Clogger represents the technological perfection of the institutional pathology Lessig has diagnosed for two decades: the capacity of concentrated interests to shape political outcomes through mechanisms that operate below the threshold of democratic accountability.

The essay distinguishes Clogger from conventional political advertising along three dimensions. First, scale: the system could generate millions of individually tailored messages, each optimized for the specific psychological profile of the recipient. Second, adaptation: the system could learn from responses in real time, adjusting its persuasive strategies based on what worked and what failed. Third, invisibility: unlike a television ad, which announces itself as political communication, Clogger's outputs would be indistinguishable from organic content — social media posts, forum responses, apparent personal communications from friends or trusted sources.

Fung and Lessig's conclusion was not that AI-enabled persuasion should be banned — they acknowledged both the difficulty of defining the prohibited behavior and the legitimate uses of AI in political communication. Their argument was for transparency requirements and for structural limits on the scale and adaptiveness of persuasive systems in political contexts. The essay has been cited in subsequent policy proposals in the United States, Europe, and elsewhere.

The thought experiment intersects with Fung's broader work on empowered participatory governance. The December 2024 Ash Center workshop on AI and democracy movements documented how AI has created asymmetric advantages for authoritarian governments over civil society, accelerating the decline of democracy movements globally. Clogger is the electoral corollary of this structural asymmetry.

Origin

Archon Fung and Lawrence Lessig published 'How AI Threatens Democracy' in The Conversation on June 12, 2023. The essay introduced the Clogger thought experiment and argued for immediate policy attention. Fung and Lessig subsequently developed the analysis in workshops at Harvard's Ash Center for Democratic Governance and Innovation, where Fung serves as director and Lessig as affiliated faculty. The thought experiment has been incorporated into multiple policy frameworks, including provisions of the EU AI Act addressing AI in political communication.

Key Ideas

Every component already exists. Clogger is not speculative technology; it is an obvious application of existing capabilities.

The essay distinguishes Clogger from conventional political advertising along three dimensions

Scale, adaptation, invisibility. The three features that distinguish AI-enabled persuasion from conventional political communication.

The governance question is temporal. Democratic deliberation about the framework must precede the deployment, not follow it.

Structural asymmetry favors autocrats. AI capabilities are more useful for concentrated power than for distributed civic action.

Transparency and structural limits. The proposed response is not prohibition but architectural constraints on scale and adaptiveness in political contexts.

In The You On AI Encyclopedia

Origin

Key Ideas

Related Entries

Further Reading