New Harvard Business School Research Takes Aim At Some Of AI’s Biggest Assumptions

Harvard Business School faculty are out this week with a cluster of new research papers and case studies that land squarely in some of the biggest debates in business right now: AI hype versus reality, the growing influence of AI-generated financial news, geopolitical fragmentation, and the limits of globalization.

Among the most topical papers is a revised HBS working paper that takes a skeptical but practical look at the use of large language models for market research. In “Using LLMs for Market Research,” professors James Brand, Ayelet Israeli, and Donald Ngwe conclude that AI tools can sometimes approximate human survey responses – but can also produce inaccurate or even “wrong-signed” results when predicting consumer preferences.

The authors argue that LLMs are best used as a supplement to human research, not a replacement for it. They found the strongest results came when models were fine-tuned using prior human survey data from similar product categories and customer groups.

AI MAY CHANGE HOW INVESTORS READ THE NEWS

Another new HBS working paper explores how AI-generated summaries are already reshaping investor behavior.

In “Generative AI and Investor Processing of Financial Media,” professors Tony Cho, Allen Huang, Joseph Pacelli, and Kristina Rennekamp analyze the rollout of AI summaries in The Wall Street Journal and find that articles with AI-generated summaries trigger higher trading volume and stronger market reactions.

The researchers also challenge a common fear about AI summaries – that they encourage shallow reading and skimming. Instead, their experiments suggest the summaries may actually improve comprehension and recall of full articles while directing investor attention toward highlighted information.

The study concludes that AI-generated summaries are becoming “interpretive cues” that shape how investors consume and react to financial news.

A NATIONAL AI STRATEGY FOR GREECE

HBS also released a new case study examining how governments may try to use AI as a national economic strategy.

“A Great Greek Leap? AI as a Catalyst for Education, Entrepreneurship, and National Prosperity,” by George Serafeim, Debora Spar, and Nicole Zelazko, centers on a partnership between Greece and OpenAI to introduce ChatGPT Edu into Greek secondary schools and launch a startup accelerator backed by OpenAI technology and mentorship.

The case frames AI not simply as a productivity tool, but as a potential lever for reversing brain drain, modernizing education, and accelerating entrepreneurship in countries still trying to rebuild after years of economic instability.

QUESTIONS ABOUT AMERICA’S FINANCIAL DOMINANCE

Another timely paper comes from professors Wenxin Du, Ritt Keerati, and Jesse Schreger, who argue that the long-assumed connection between U.S. Treasury dominance and dollar dominance may be weakening.

Their forthcoming IMF Economic Review paper, “Decoupling Treasury and Dollar Privilege,” finds that while the U.S. dollar has retained its global “convenience” and safe-haven status since the global financial crisis, the relative attractiveness of U.S. Treasuries has deteriorated – even turning negative at some maturities compared with other developed-market government bonds.

The paper arrives as economists and policymakers increasingly debate the long-term durability of U.S. financial leadership amid rising debt levels and geopolitical fragmentation.

WHY RETAILERS FAIL ABROAD

Meanwhile, another HBS working paper argues that many retailers misunderstand why international expansion efforts fail.

In “Retail Expansion to International Markets: Why Some Retailers Succeed and Many Fail,” professors Srikant Gokhale and Rajiv Lal contend that the biggest problem is not localization mistakes after entering a market, but failing beforehand to identify what parts of a company’s business model are essential and which can be adapted.

“The strategic error is not poor localisation,” the authors write. “It is an absence of balanced adaptation.”

DON’T MISS HARVARD LAUNCHES ‘AI FOUNDRY,’ MAKING THE CASE FOR MOVING BEYOND THE CASE STUDY

© Copyright 2026 Poets & Quants. All rights reserved. This article may not be republished, rewritten or otherwise distributed without written permission. To reprint or license this article or any content from Poets & Quants, please submit your request HERE.