The Shrinking Synthesis: Information-Technology Settlement Cycles and the 2037–2047 Window for AI’s Institutional Reformation

Five Western information-technology cycles form the empirical core: the printing press (settlement time 249 years), the telegraph and mass press (46), radio (28), television (20) and the internet (23). Each runs the same five-phase shape — diffusion, capture, chaos, foundational institutional response, consolidation — and each is measured to a discrete, datable Phase 4 instrument: the English Bill of Rights, the Sherman Antitrust Act, the UN Declaration of Human Rights, the ICCPR, the GDPR.

The compression curve flattens at roughly a generation across the three most recent cycles, and the paper proposes a mechanism for the floor: institutional adaptation cannot compress below the time required for a cohort to grow up with the technology. The argument stands on three legs — cohort replacement (Mannheim, Ryder), organisational learning (Argyris and Schön; Levitt and March) and long-horizon institutional formation (Ostrom, Brand). Settlement waits not for the technology to mature but for the people formed by it to reach the rooms where institutions are written.

Applied prospectively to AI, with ChatGPT’s November 2022 release as diffusion onset, the framework predicts a foundational institutional response in the 2037–2047 window, central estimate around 2042. The settlement is expected to be multipolar: each major tradition producing its own accommodation — individual-rights claims in the Anglophone democracies, distributional claims in the European social democracies, social-stability claims in China’s institutional orbit — convergent in timing, divergent in content.

Three predictions are deposited on the Open Science Framework with vindication and refutation conditions stated in advance: the timing window, the cohort that will author the settlement, and the conceptual lineage its content will carry. A criterion-meeting event before 2037 disconfirms the generational floor; none by 2049 disconfirms the compression pattern. Either refutation would be informative, which is the point of registering them.