Matthew DosSantos DiSorbo

My research centers on identifying — and eliminating — bottlenecks in consumer-facing AI systems. Specifically, my work explores two streams: 1) how the framing of AI recommendations and design of AI systems influence judgement and prediction at the consumer-AI interface, and 2) how hidden patterns in AI reasoning and decision-making impact autonomous service systems.

I will receive my PhD in Technology and Operations Management Unit at Harvard Business School in May 2026 under the advisement of Kris Johnson Ferreira. I will be presenting at the 2026 MSOM SIG Day (Sunday, 7/12).

Research

Publications

• Warnings and Endorsements: Improving Human-AI Collaboration in the Presence of Outliers
  With Kris Johnson Ferreira, Maya Balakrishnan and Jordan Tong. Manufacturing & Service Operations Management.
  Winner, 2024 Production and Operations Management College of Behavioral Operations Management Junior Scholar Paper Competition (INFORMS)
  Media: How Decision-Makers Can Catch Gen AI's Bad Advice (HBS Working Knowledge)
  
  
• Teaching AI to Handle Exceptions: Supervised Fine-Tuning with Human-Aligned Judgment
  With Harang Ju and Sinan Aral. PNAS Nexus.
  Media: 4 new studies about agentic AI from the MIT Initiative on the Digital Economy (MIT Sloan)
  
  

Working Papers

• Accuracy Obsession: Humans Prioritize Immaterial AI Accuracy Over Their Own Compensation — Unless We Educate Them
  With Kris Johnson Ferreira. Accepted for presentation at the 2026 MSOM SIG Day.

• The GenAI Wall Effect: Examining the Limits to Horizontal Expertise Transfer Between Occupational Insiders and Outsiders
  With Luca Vendraminelli (first author), Annika Hildebrandt, Edward McFowland III, Arvind Karunakaran and Iavor Bojinov.
  Media: Will AI Make Jobs Less Specialized? (Time.com); Hitting the ‘GenAI wall’: Where generative AI stops working, and what it means for your talent strategy (Fortune.com)
  

• Act or Escalate? Evaluating Escalation Behavior in Automation with Language Models
  With Harang Ju. Submitted to the 2026 Conference on Language Modeling.
  

• Evaluations Amid Measurement Error: Determining the Optimal Timing for Workplace Interventions
  With Iavor Bojinov and Fiammetta Menchetti. Major Revision at the American Journal of Evaluation.
  
  
• I Committed It to Memory but Then I Forgot: Chat Template Tokens Alone Break LLM Composition of Consumer Information
  With Harang Ju. Work in progress.
  
  

Teaching

Harvard Business School

• Data Science for Managers — Required Curriculum, Fall 2023
• Data Science for Managers — First Year Elective, Fall 2022

Harvard University

• Introduction to Probability — Fall 2023, Fall 2016, Fall 2015
• Introduction to Stochastic Processes — Spring 2017
• Introduction to Statistical Inference — Spring 2016
• Introduction to Quantitative Methods for Economics — Spring 2015, Fall 2014

Awards

Certificate of Special Recognition from the Harvard Undergraduate Association (Fall 2023),
Derek Bok Certificate of Distinction in Teaching (6× winner, Fall 2014–Spring 2017),
2016 David Pickard Teaching Fellow in the Harvard Statistics Department.

Student Reviews

Full reviews can be found here.

I am in my fourth year as a Resident Tutor in Winthrop House, one of twelve undergraduate Houses at Harvard College. Below are selected responses to "How have you benefited from interacting with your Resident Tutor?" from the Harvard College Residential Life Survey.