Artificial intelligence is rapidly dismantling longstanding barriers related to time, labor and data availability in economic research, according to a new column by Isabel Vansteenkiste, a European Central Bank official, published by the Centre for Economic Policy Research (CEPR). However, while AI is accelerating research output and the dissemination of knowledge, it could also exacerbate existing institutional bottlenecks across the economics profession.
Validation bottlenecks and lost training opportunities
As AI sharply reduces the marginal cost of producing working papers, Vansteenkiste warns that the burden of peer review and formal validation will grow significantly. Journals and readers may face a flood of plausible but potentially inaccurate automated findings, increasing reliance on reputational shortcuts and widening disparities between elite and under-resourced research institutions.
Read also: Why philosophers are being hired to train AI models.
Vansteenkiste also notes that as AI takes over routine tasks such as data cleaning and basic literature reviews, junior researchers risk losing the hands-on experience that traditionally helps develop critical analytical skills.
From Galileo’s answers to Socratic questioning
Beyond these procedural challenges, AI could fundamentally reshape the scientific process that evolved during the Renaissance. Traditional research seeks to uncover underlying causal mechanisms, while advanced AI models often prioritize predictive accuracy without providing conceptual explanations. This creates an epistemic gap between automated empirical predictions and deeper scientific understanding.
Ultimately, economic research may need to shift away from Galileo’s historical emphasis on finding definitive answers and return to Socrates’ enduring focus on asking incisive questions. In an era increasingly shaped by confident machine-generated outputs, human scholars must preserve their capacity for critical doubt, rigorous institutional oversight and methodological transparency to safeguard the long-term integrity of scientific research.