Managing the Transition to an Economy of Freely Chosen Activity under Pervasive General-Purpose Robotic Automation: A Mathematical and Agent-Based Policy Model with a Critical Review of the Literature
We develop a federal-level decision-support framework for managing a long-run transition to an economy in which artificial intelligence and inexpensive general-purpose robots can perform a substantial share of cognitive and physical tasks. The framework does not forecast the technology's arrival: 2041, 2046, and 2051 serve as conditional midpoint scenarios, and one policy package must perform acceptably across all of them. A discrete-time mathematical system is coupled with an agent-based simulation of heterogeneous citizens, six sectors, and a dedicated public transition account. The policy vector combines a tax on automation rents, statutory public participation in robotic capital, a social dividend, lifelong-learning capacity, a public fund, and staged work-time reduction. An operational index measures the capacity for free activity: voluntarily usable time conditional on material security, skills, motivation, and trust. In the illustrative central experiment, the managed package reduces the below-threshold income share from 23.3% to 7.9% and the income Gini from 0.383 to 0.183, at the cost of 6.4% lower output per capita in 2061; a cash-dividend-only policy improves distribution but barely moves free-activity capacity. No numerical level is an estimate for any real economy: the experiment compares mechanisms within one illustrative parameterization. We provide a pre-specified calibration protocol (method of simulated moments, backtesting, ensemble re-optimization), a staged roadmap with monitoring indicators and stopping rules, and a critical review of six literature strands. Code, random seeds, tables, and figures are fully reproducible. The deposit includes the full replication package: the agent-based model (Python/NumPy), experiment driver, automated tests with bit-level reproducibility checks, machine-readable results, figures, and the deterministic Markdown-to-LaTeX build pipeline. See README_EN.md inside the archive.