Economic research pivots to data-driven methods amid reproducibility crisis, AI hype, and entrenched institutional incentives
Original framing: “How economics is changing” — Financial Times
Reproducibility audits in top economics journals show that less than 40% of published empirical findings survive replication attempts, a figure comparable to the biomedical crisis. AI models exacerbate this problem when training data are proprietary and opaque, hindering peer verification. Strengthening open data mandates and pre‑registration can restore scientific credibility.
The transformation of economics into a data‑driven, AI‑infused discipline is not merely a technical upgrade but a systemic shift that reproduces longstanding power imbalances unless deliberately restructured.
Historical cycles of methodological hype, coupled with entrenched academic incentives, have generated a reproducibility crisis that erodes public trust. By foregrounding Indigenous stewardship, cross‑cultural policy experiments, and marginalised scholarly voices, we can redesign AI tools to serve equitable, ecologically sound futures. Institutional mechanisms—open‑science mandates, public AI oversight, pluralist epistemic integration, and incentive reform—must be coordinated across universities, funding bodies, and international regulators to prevent the new wave of technocratic dominance from eclipsing diverse, resilient economic imaginaries.