S9 Ep2: Has AI eaten the economics major?

Episode Summary

Executive Summary: Wendy Carlin argues that AI should force a rethink of economics teaching, not just assessment. CORE replaces model-first, lecture-led teaching with a world-first approach built around real problems, critical reading, and active learning. She says Gen AI can help with routine problem sets, but economics education must shift toward curiosity, ambiguity, and judgment.

Main Topics: Why CORE was created (Priority: 5/5): CORE emerged after the financial crisis from concern that undergraduate economics lagged behind research and real-world problems, especially in what students were taught and what employers needed. How CORE teaches economics differently (Priority: 5/5): Instead of starting with standard models and diagrams, CORE begins with real-world facts and cases, then uses models to explain and critique them. Evidence on CORE outcomes (Priority: 4/5): Carlin cites a quasi-experimental study suggesting students taught with CORE performed better in later courses, implying stronger learning skills beyond the economics classroom. Gen AI’s impact on problem sets and learning (Priority: 5/5): She argues AI is very good at structured learning tasks and can support tutoring, but also threatens traditional assignments and encourages superficial completion of work. Employer demands and graduate skill gaps (Priority: 5/5): She says long-standing employer complaints about economics graduates—especially weak writing, teamwork, and real-world application—are now amplified by AI. What should change in economics curricula (Priority: 4/5): Carlin advocates reducing repetitive problem-set sessions, redesigning pedagogy around research-like inquiry, and using experiments and AI tools as teaching aids. CORE’s global adoption and collaboration model (Priority: 3/5): CORE has spread through voluntary, bottom-up collaboration and translations, with a large international user base and ongoing work to expand access.

Key Arguments: Gen AI can already complete many standard economics problem sets, so routine assessment is no longer a sufficient measure of student learning. A better use of AI is as an adaptive tutor that accelerates advanced learners and supports struggling learners through repetition and interleaving. CORE’s pedagogy teaches students to start from real-world phenomena, build models, and then evaluate what the model explains and misses. Students need skills that AI cannot easily replicate: independent reading, critical evaluation, handling ambiguity, and applying models to messy real-world problems. Employer surveys have long shown that economics graduates are strong quantitatively but weak in writing, teamwork, cross-cultural sensitivity, and critical application. Academics have been slow to change because incentives reward research over teaching innovation, and switching costs for existing courses are high. Economics teaching should become more like economics research: framed around questions, data, and models rather than fixed textbook routines. AI can be turned into a cognitive sparring partner, exposing simplistic reasoning and prompting deeper instructor-led discussion.

Data Points: Countries using CORE: about 80 - Global adoption of CORE materials worldwide CORE study effect size: 0.26 standard deviations - La Trobe Business School study found students taught with CORE performed better in later courses Initial CORE meeting year: 2013 - First meeting organized by Carlin to assess whether economics teaching needed change Number of people at initial meeting: about a dozen - Early discussion on whether undergraduate economics needed reform CORE translation status: 3 languages mentioned - Latest flagship The Economy 2.0 being translated into Chinese and Korean, with Spanish completed Problem set sessions: 20–30 students for 1 hour - Traditional TA-led sessions described as inefficient and now a waste of time Experiencing Economics time: 10 minutes - Suggested classroom engagement time for CORE experiments CORE content horizon: last 30–40 years - Curriculum updated to incorporate newer research beyond Marshall plus Keynes Survey start year: 2012 - Royal Economic Society employer survey referenced as evidence of long-standing skill gaps

Pivotal Quotes: "we do not need to have 20, 30 students sitting in a room for an hour with a TA going through a problem set, which was always pretty much a waste of time. And it's now a complete waste of time." — Wendy Carlin: On why Gen AI and new pedagogy make traditional problem-set sessions less valuable "we want to turn it into a cognitive sparring partner." — Wendy Carlin: Describing the desired role of AI in economics teaching "The worst outcome is certainly downgrading of students' ability to think independently and to read independently" — Wendy Carlin: On the main risk if economics teaching does not adapt to Gen AI

Implications: Economics departments should redesign teaching around critical thinking, real-world inquiry, and AI-aware assessment. Routine problem sets will matter less; independent reading, judgment, and interpretation will matter more.