The Deployment Question
The deployment question is this book's final framing of the AI challenge: toward what ends will the second cognitive surplus be directed, by whom, and for whose benefit? The question is not answered by the technology. Every surplus in human history has presented the same question. The Agricultural Revolution produced a surplus of food that could have been distributed broadly but for most of twelve thousand years was captured by elites. The Industrial Revolution produced a surplus of goods that could have been distributed broadly from the beginning but for the first century was captured by factory owners. The first cognitive surplus produced a surplus of participation that could have been deployed entirely toward civic and collective purposes but was in significant part captured by platforms whose business model depended on engagement rather than value. The historical pattern is unambiguous: the default deployment of a surplus is capture by the actors best positioned to capture it, not distribution toward the actors who would benefit most from it.
In the AI Story
The actors best positioned to capture the second surplus are the platform companies providing the AI tools. They control the means of creation, the infrastructure through which creations are shared and discovered, and the data generated by the creative process — data that constitutes the most detailed map of human need ever produced. The actors who would benefit most from broad deployment are the ones least visible in current discourse: communities whose needs are unaddressed by commercial software, populations whose creative potential is constrained by barriers of access and connectivity, public institutions that serve populations too small or too poor to attract commercial attention.
Deployment toward broad civic value rather than narrow commercial capture requires institutional action at multiple levels. Platform level: design choices that prioritize sharing, collaboration, and discovery alongside individual creation. Governance level: tiered quality systems, liability frameworks, intellectual property structures, and platform accountability mechanisms. Educational level: curricular transformation that shifts from teaching answers to teaching questions, from execution skills to judgment. National level: recognition that the nations building the best institutional infrastructure for channeling AI-enabled creation will lead the next century — not because they have the most powerful AI but because their citizens will be most capable of directing AI toward human flourishing.
National strategies for AI have focused overwhelmingly on the supply side: investing in AI research, supporting AI companies, regulating AI development. The demand side — preparing citizens to use AI tools wisely, building the institutional infrastructure that channels creative surplus toward public value, ensuring equitable access to the means of creation — remains almost entirely unaddressed. The imbalance is a predictable consequence of who has political visibility in technology policy debates: companies have lobbyists, citizens do not. But the imbalance determines the deployment, and the deployment determines whether the surplus produces democratic empowerment or concentrated capture.
The question is not determined. It is being written, right now, by every policy decision, every platform design choice, every educational reform, every governance experiment, every community forming around the shared purpose of building things that serve more than the builder. The historical pattern is unambiguous about where defaults lead. The question is whether the people, institutions, and societies affected by the surplus will act to change the defaults before the defaults harden into permanent arrangements.
Origin
The framing of deployment as the structural question synthesizes Shirky's institutional analysis with the historical pattern that earlier surpluses have established. The argument that technology does not determine deployment runs through Shirky's work from Here Comes Everybody forward and appears in parallel form in work by Carlota Perez, Daron Acemoglu, and others.
Key Ideas
Technology does not determine deployment. Every previous surplus has been deployed through institutional choices; AI will not be exceptional.
The default is capture. Without institutional intervention, surpluses are captured by the actors best positioned to capture them, not distributed to those who would benefit most.
Multi-level response required. Deployment toward broad value requires action at platform, governance, educational, and national levels simultaneously.
Supply-side bias. National AI strategies have focused on AI development while neglecting the institutional infrastructure that determines how AI is used.
The closing window. Deployment arrangements harden over time; the window for institutional intervention is narrower than the pace of capability development suggests.
Appears in the Orange Pill Cycle
Further reading
- Carlota Perez, Technological Revolutions and Financial Capital (Edward Elgar, 2002)
- Daron Acemoglu and Simon Johnson, Power and Progress (PublicAffairs, 2023)
- Shoshana Zuboff, The Age of Surveillance Capitalism (PublicAffairs, 2019)
- Al Gore, The Assault on Reason (Penguin, 2007)