AI Industrial Capture

Definition

AI industrial capture is the concentration of consequential AI decisions inside a small network of leading companies and government actors, with limited legislative, public, or independent technical scrutiny.

Current Synthesis

The source treats capture as a structural allocation-of-authority problem rather than proof of bad intent by one company. Frontier labs possess models, expertise, and infrastructure; governments possess procurement, security, and coercive authority. When the two negotiate deployment boundaries privately, practical governance can emerge before democratic institutions have established who may decide, what evidence outsiders may inspect, or how affected publics can challenge the result.

Key Claims

• Concentrated capability can turn a small number of executives and officials into de facto AI governors.
• Public-private cooperation can be necessary while still producing an accountability deficit.
• Defense procurement is a high-salience capture risk because commercial negotiations can settle warfare boundaries before public law does.
• Independent evidence access, legislative scrutiny, disclosure, and public deliberation are complementary checks on capture.
• The remedy is institutional redesign, not assigning every governance failure to one firm or individual.

Evidence

• Concentrated authority - EP 35: Who Actually Controls AI? The Governance Gap Explained says a small group of companies and CEOs make choices with civilizational implications while no international authority has effective control.
• Closed defense channel - EP 35: Who Actually Controls AI? The Governance Gap Explained uses negotiations involving Anthropic and U.S. defense officials to argue that military-AI boundaries were discussed without adequate public, congressional, or independent technical review.
• Structural diagnosis - EP 35: Who Actually Controls AI? The Governance Gap Explained explicitly frames the problem as institutional rather than as the failure of one company.

Counterevidence & Qualifications

The supplied summary does not document the underlying defense negotiations or show that participating actors excluded every formal oversight channel. Governments may need confidential technical and national-security discussions, and frontier labs may hold expertise that public bodies lack. The concept identifies a legitimacy and accountability risk; it does not establish corrupt intent or prove that every public-private AI decision is captured.

What Changed

• Established a public-private concentration concept distinct from narrower regulatory capture by incumbents.
• Identified defense contracting as a pathway through which deployment rules can precede democratic authorization.

Related Concepts

• AI Regulatory Capture Risk - narrower risk that formal safety rules protect leading firms or become incumbent moats.
• Defense AI Procurement - contracting pathway where public and private deployment authority meet.
• Democratic AI Governance Deliberation - public-legitimacy response to closed decision channels.
• Frontier Model Disclosure - evidence-access mechanism that makes external scrutiny more feasible.
• AI Alignment Governance - broader question of aligning institutions and the people who govern AI builders.