Dead Economy Theory: AI Displacement and the Erosion of Democratic Stability
Owen McGrann’s latest piece argues that ‘so-so’ automation is displacing labour without generating productivity, threatening the economic leverage that underpins democratic systems.
An analysis published on 29 May 2026 by Owen McGrann argues that the rapid adoption of artificial intelligence by major technology firms poses a systemic risk to democratic governance. The article contends that AI-driven labour replacement creates a 'dead economy' where consumer demand collapses as workers lose income, while wealth concentrates among corporate owners. Citing economists such as Daron Acemoglu and Brett Hemenway Falk, the piece suggests that 'so-so' automation displaces jobs without generating significant productivity gains, leading to social instability and the erosion of the political leverage held by the general populace.
McGrann’s analysis highlights the vast capital investments by firms such as OpenAI, Anthropic, and Google DeepMind, noting that valuations in the hundreds of billions require a market size only comparable to the global labour market. The piece argues that investor presentations framing AI as a 'copilot' are marketing cover for a financial model predicated on the elimination of human cost centres. With benchmarks showing AI models achieving high win rates against human professionals, the industry is explicitly targeting the professional class for displacement.
The article outlines a cyclical economic trap where firms replace workers to boost margins and stock prices, only to subsequently lose those workers as consumers. Citing research from Wharton economists Brett Hemenway Falk and Gerry Tsoukalas, the analysis describes a 'prisoners’ dilemma' where competitive pressures force firms to automate beyond the socially optimal level, destroying the aggregate demand required to sustain the economy. This dynamic is exacerbated by herd behaviour, where executives cut jobs in anticipation of competitors’ moves, regardless of immediate efficiency gains.
Historical comparisons are drawn to the agricultural and industrial transitions, which took decades or centuries to stabilise wages and employment. McGrann contrasts this with the speed of AI adoption, citing former National Economic Council deputy director Bharat Ramamurti, who warns the displacement could unfold in years rather than decades. The piece notes that unlike previous narrow technologies, general-purpose AI threatens cognitive labour across all industries simultaneously, with economist Daron Acemoglu observing that the displacement effect of new technologies currently outweighs productivity gains.
The analysis concludes that the erosion of labour as a production input undermines the material basis of democratic leverage. With tax bases shrinking and collective bargaining becoming vestigial, wealth concentration accelerates, potentially leading to social instability. McGrann points to comments from Anthropic CEO Dario Amodei, who has acknowledged that democracy relies on the average person having economic leverage, a condition the technology he is building may remove. The piece warns that without regulatory intervention, the economic incentives for AI companies may align more closely with authoritarian regimes that face no electoral constraints on workforce displacement.
