Market Design in the Age of AI, this Friday at Stanford

 Market Design in the Age of AI
Friday, February 27, 2026 , 12:00pm - 6:00pm PST
Simonyi Conference Center, CoDa, 389 Jane Stanford Way, Stanford, CA 94305

The Market Design in the Age of AI Conference aims to catalyze interdisciplinary collaboration and innovation at the intersection of algorithm design, economics, machine learning, and operations research. As market platforms become increasingly complex and data-driven, this conference brings together leading thinkers from both academia and industry to explore how AI is reshaping the design, analysis, and regulation of modern markets.

Our goal is to foster a vibrant community that bridges research and practice—advancing both the theory and real-world application of intelligent, efficient, and equitable market systems. Over time, we envision the conference serving as a launch pad for novel marketplaces and platform innovations that emerge from these collaborations.

Agenda (draft)

Start Time	End Time	Session	Speaker(s)
12:00 PM	1:00 PM	Lunch & Registration
1:00 PM	1:15 PM	Opening Remarks	Amin Saberi, Professor of Management Science & Engineering, Stanford; Computational Market Design Center Director
1:15 PM	1:45 PM	Economic Mechanisms in the GenAI Era: Advertising Auctions and Marketplaces	Aranyak Mehta, Distinguished Research Scientist, Google
1:45 PM	2:15 PM	Carpooling and the Economics of Self-Driving Cars	Michael Ostrovsky, Professor of Economics, Stanford Graduate School of Business
2:15 PM	3:00 PM	GenAI for Markets	Amine Allouah, My Custom AI, Custom AI James Biggs, Sr. Director of Marketplace Ads & Monetization Products, Upwork Yossi Feinberg, Professor of Economics, Stanford Graduate School of Business Negin (Nikki) Golrezaei, Theresa Seley Associate Professor in Management Science, MIT Ruben Lobel, Product Data Scientist Lead, Waymo Okke Schrijvers, Research Scientist Manager, Meta
3:00 PM	3:15 PM	Break
3:15 PM	3:45 PM	TBA	Vahab Mirrokni, VP & Fellow, Google Research
3:45 PM	4:15 PM	Fireside Chat: The Economics of AI	Michael Schwarz, Corporate Vice President & Chief Economist, Microsoft Guido Imbens, Professor, Stanford University; Faculty Director, Stanford Data Science Amin Saberi, Professor of Management Science & Engineering, Stanford; Computational Market Design Center Director
4:15 PM	4:45 PM	TBA	Ronnie Chatterji, OpenAI Chief Economist; Duke University Professor of Business & Public Policy
4:45 PM	5:00 PM	Closing Remarks
5:00 PM	6:20 PM	Networking Reception & Posters	View the Poster Map

 

Organizers

• Negin (Nikki) Golrezaei, MIT/Stanford
• Amin Saberi, Stanford

Poster Session Committee

• Mohammad Asadi, HAI Graduate Fellow
• Grace Guan, SDS Alumna (PhD in Management Science and Engineering)
• Andy Haupt, HAI Postdoctoral Fellow

A.I. in managemant consulting

 Management consulting seems like a natural use-case for large language models.

The Financial Times has the story: 

Accenture combats AI refuseniks by linking promotions to log-ins
Consulting firms use ‘carrot and stick’ with some senior staff less willing to use technology than junior colleagues     by Ellesheva Kissin and Elizabeth Bratton 

"Accenture has begun monitoring staff use of its AI tools as part of how it decides top-level promotions, as consultancies push reluctant employees to adopt the technology. 

The Dublin-headquartered firm told associate directors and senior managers that promotion to leadership positions would require “regular adoption” of AI, according to people familiar with the matter and an internal email seen by the FT.

This month Accenture started to collect data on individual weekly log-ins to its AI tools for some senior employees."

Fast and slow dissemination of new ideas in medicine and economics (one timeline:)

There are many differences between medicine and economics, but one of the most striking is the speed of publication. 

I publish papers in both fields, so I get to experience very different speeds, of publication and response. Publishing (and therefore also responding--both positively and negatively ) is much slower in economics than in medicine.  I've been noticing this because of recent attention to a paper I coauthored that was published in November, 2025, in the journal Transplantation. (It had been submitted in January, was revised and accepted in March, and was published online in May.)  In December the journal created and distributed to its subscriber list a narrated video abstract of the paper. You can find the video here https://vimeo.com/1146995735/486989e95c?fl=pl&fe=sh

 Our paper suggested ways that information revealed during deceased organ allocation could be used to evaluate organ quality, and expedite (i.e. speed up) the allocation process for organs at risk of being unused.  And the first published response, just three month later, suggests how such information could be used in India.

Early Refusal Pattern Phenotyping as a Surrogate for Organ Quality Assessment in Kidney Allocation
Kashiv, Pranjal MD, DM1,2; Pasari, Amit MD, DM2,3; Balwani, Manish MD, DM2,3; Kute, Vivek MD, DM4
Transplantation ():10.1097/TP.0000000000005664, February 09, 2026. | DOI: 10.1097/TP.0000000000005664

"We read with interest the recent article by Guan et al,1 which provides a comprehensive and methodologically thoughtful assessment of refusal behavior in deceased donor kidney allocation. Their distinction between single-patient and multiple-patient simultaneous refusals, derived through timestamp-based clustering, offers a methodologically robust framework that elevates routine offer-response data into a meaningful surrogate for real-time assessment of organ suitability. This approach is particularly valuable in allocation environments where decisions must be made under substantial time pressure and with incomplete ancillary information.

...

" Their observations offer global relevance and hold potential for strengthening allocation efficiency in India’s evolving deceased donor landscape." 

 ######

Earlier:

Here's the blog post that accompanied the publication online... 

Friday, May 23, 2025  Deceased organ allocation: deciding early when to move fast

French medical residents struggle without a way for couples to match together

 Esther Duflo sent me a link to the story below in Le Monde, describing the difficulty facing married doctors in France who wish to match to first jobs in the same city.

If I understand correctly, the old system ranked students by their exam results on the  épreuves classantes nationales (ECN), after which the students  then chose among the jobs available to them in rank order. A couple could decide to choose jobs at the same time, by waiting till the turn of the worse ranked member of the couple.  In this way they could choose two jobs in the same city.  But a new system ranks doctors differently, by specialties, and resolves itself for everyone at the same time. So couples can now have no way to assure themselves of being matched in the same city. 

Here's the article (which I sense isn't completely clear either in English translation or in the original French): 

"We are pawns to be moved": the geographical conundrum of medical intern couples
Since the last reform of medical studies, the algorithm that distributes future interns in the university hospitals of France no longer guarantees their assignment in the same city as their partner.  By Diane Merveilleux 

" Since 2023, the three-day ECNs giving rise to a single national ranking have been replaced by a sixth year punctuated by tests in the form of multiple-choice questions (MCQs) in October, then oral exams in the spring, which make it possible to establish 13 rankings by specialty groups.

"When summer comes, each student makes up to 80 wishes, a wish corresponding to a specialty in a city. Then an algorithm makes it possible to assign each person an assignment according to the ranking of their wishes. This algorithm will run eight times, each time simulating a final result and allowing students to adapt their wishes and review their ambitions. Their final assignment falls at the same time for everyone, at the end of the last round, in September. They are then hired for four to six years of boarding in the same place, without any possibility of change.

"For young medical couples, this is where the problem lies: those who would like to spend these years in the same city have no way of being sure.

...

"Thus, until 2023, the national ranking led students to choose their assignment one after the other – the first, then the second, and so on. In this configuration, the member of the couple who had obtained the highest score could "downgrade" themselves to join their partner's level so that they could make their decision at the same time. "Today, this downgrading is no longer possible, it's the very principle of the algorithm: we have a system that is totally different," says Philippe Touzy."

########## 

It seems that the couples algorithm that they have abandoned is much more primitive than the Roth-Peranson algorithm* now used in the American medical match, which allows couples to submit much more expressive joint preferences over pairs of jobs.

As it happens, just a few years ago there was a proposal for a two-part match that would have inadvertently disabled the US couples match.  Itai Ashlagi and I were able to convince the American Medical Association (which had sponsored the proposal) that it was a bad idea.  We ran the match under a variety of conditions.  So Itai  has an efficient implementation of the Roth-Peranson algorithm available if it could be helpful in France.

Here’s a blog post about that…

Friday, April 21, 2023

Transition from medical school to residency: defending the parts that work well (namely the NRMP Resident Match)

########

* 

 From     Roth, Alvin E. "What have we learned from market design?" Reprinted with a postscript in The Handbook of Market Design, Nir Vulkan, Alvin E. Roth and Zvika Neeman editors, Oxford University Press, 2013, Chapter 1, 7-50

How to regulate legal marijuana?

 The New York Times editorial board thinks about the current environment for (now legal) marijuana, and calls for more careful regulation, and federal taxation:

It’s Time for America to Admit That It Has a Marijuana Problem 

"Thirteen years ago, no state allowed marijuana for recreational purposes. Today, most Americans live in a state that allows them to buy and smoke a joint. President Trump continued the trend toward legalization in December by loosening federal restrictions.

This editorial board has long supported marijuana legalization. In 2014, we published a six-part series that compared the federal marijuana ban to alcohol prohibition and argued for repeal. Much of what we wrote then holds up — but not all of it does.

At the time, supporters of legalization predicted that it would bring few downsides. In our editorials, we described marijuana addiction and dependence as “relatively minor problems.” Many advocates went further and claimed that marijuana was a harmless drug that might even bring net health benefits. They also said that legalization might not lead to greater use.

 

It is now clear that many of these predictions were wrong. Legalization has led to much more use. Surveys suggest that about 18 million people in the United States have used marijuana almost daily (or about five times a week) in recent years. That was up from around six million in 2012 and less than one million in 1992. More Americans now use marijuana daily than alcohol. 

...

"The unfortunate truth is that the loosening of marijuana policies — especially the decision to legalize pot without adequately regulating it — has led to worse outcomes than many Americans expected. It is time to acknowledge reality and change course." 

 

Magic mushrooms have a role in hospice care

 Pain experienced while dying may be partly spiritual.

 National Geographic has the story: 

These drugs could be a game changer for end-of-life care
Certain psychoactive substances can improve the mental health of terminally ill cancer patients—but few patients can currently access them.  By Meryl Davids Landau

 "Several years ago in Vancouver Island, Canada, a 32-year-old mother with advanced metastatic cancer was so wracked with pain and a fear of dying she constantly wept in bed. Through a targeted Canadian government program, the woman accessed psilocybin, the main psychedelic ingredient in magic mushrooms. The day after taking a dose of the drug she was pain-free, able to joke with family members and reconnect with old friends before she died the following week.

...
"The drugs can help with “the existential component of pain that is tied in with spiritual and psychological experiences,” something conventional medicine has few tools to address, says Masuda, a physician with SATA Centre for Conscious Living, who has since facilitated dozens of psychedelic sessions for similar patients.

"Some 400 terminal patients in Canada have legally accessed psilocybin in the past five years via its special programs, and several countries already allow for similar uses. Due to federal drug laws, terminally ill people in the U.S. cannot currently take psilocybin outside of a handful of clinical trials.

"But this may finally change, as government agencies are evaluating whether to allow its use for end-of-life care—thanks to pressure from physicians and years of research. Many palliative care doctors in the U.S. say the change can’t come soon enough." 

Jobs for human "meatspace" workers, assigned by A.I.s

 Robots aren't yet able to replace people: e.g. self-driving taxis (such as Waymo) aren't equipped to close a door left open (or incompletely closed) by a departing passenger.  So artificial agents need a task rabbit to recruit able-bodied (or at least embodied) workers.  

Nature has the story: 

AI agents are hiring human 'meatspace workers' — including some scientists
Biologists, physicists and computer scientists have joined a platform called RentAHuman.ai to advertise their skills. By Jenna Ahart 

"The idea is simple, as the website’s homepage reads: “robots need your body”. Human users can create profiles to advertise their skills for tasks that an AI tool can’t accomplish on its own — go to meetings, conduct experiments, or play instruments, for example — along with how much they expect to be paid. People — or ‘meatspace workers’ as the site calls them — can then apply to jobs posted by AI agents or wait to be contacted by one. The website shows that more than 450,000 people have offered their services on the site." 

Joe Halpern (1953-2026)

 Joseph Halpern was an early explorer of the interface between computer science and game theory.  

Here's his funeral home obit: 

Joseph Y. Halpern
May 29, 1953 — February 13, 2026 

"Joe spent nearly 30 years as a professor of computer science at Cornell, and was considered a pioneer in his field. He was famous for having an impressive influence in a wide variety of topics, working extensively at the intersection of computer science, philosophy, and game theory. His work has reshaped the way we think about topics such as reasoning about knowledge and causality. He is the recipient of prestigious awards such as the Gödel Prize and Dijkstra Prize, the co-author of three highly influential books, six patents, and over 300 papers." 

 

His student Daphne Koller writes:

In Memoriam: Joe Halpern

"Yesterday, my PhD advisor, Joe Halpern, passed away after a long battle with lung cancer. He was a brilliant mathematician, a transformative mentor, and a truly wonderful human being.

"Joe possessed the rare ability to identify unusual, deeply interesting problems and solve them with breathtaking elegance and rigor. He was also a master communicator who could distill the most complex concepts into simple, straightforward truths—a skill I strive to emulate every day." 

 

Here's his Google Scholar Page: Jospeh Halpern 

Eric Schmidt on the future of warfare

 Warfare--the technology by which wars are fought--is changing.

Ukraine’s no man’s land is the future of war
In drone vs drone combat, valuable personnel can be pulled back from the front
Eric Schmidt 

"The writer is former CEO of Google, chair of the Special Competitive Studies Project and an investor in drone technology" 

"Future wars are going to be defined by unmanned weapons. The combination of unblockable satellite communications, cheap spectrum networks and accurate GPS targeting means the only way to fight will be through drone vs drone combat. Drones share data in real time, meaning that many inexpensive platforms can act as a single weapon. They will carry air-to-air missiles to defeat attackers, just like a fighter jet does, but will be cheaper and more abundant.

"The winner of those drone battles will then be able to advance with unmanned ground and maritime vehicles, which move slowly but can carry heavier payloads. These air, land and sea formations will absorb the initial fire and expand what is becoming an increasingly robotic kill zone. Only after the first waves of machines have gone in will human soldiers follow."
 

4th Computational and Experimental Economics Summer School, UPF Barcelona, May31-June 6

Rosemarie Nagel writes from Barcelona:

Dear colleagues and graduate students,

 We invite graduate students, postdocs, and young faculty to the 4th Computational and Experimental Economics Summer School, to be held on May 31 - June 6, 2026, at the BESLab at UPF in Barcelona, Spain.

 The goal of the summer school is to build a foundation for using computational tools, machine learning methods, and large language models to complement and/or explain results from human-subject experiments. In particular, throughout the curriculum, students will learn to implement a variety of agent-based models that have successfully captured regularities observed in experimental and field data.  

 

In addition, the summer school will include a two-day workshop on computational and experimental economics at the BSE summerforum, featuring presentations by leading researchers in experimental and computational economics.

 The deadline for applications is March 7th, 2026, for the summer school. 

 You can find more information and details on how to participate in the summer school here: https://www.upf.edu/web/beslab/comp-2026

 Organizers,

 Herbert Dawid (Bielefeld University)

Mikhail Anufriev (University of Technology Sydney)

Rosemarie Nagel (ICREA-UPF, and BSE)

Valentyn Panchenko (University of New South Wales)

Yaroslav Rosokha (Purdue University)

Trump Administration Removes Pride Flag From Stonewall National Monument (but the Stonewall Inn is still in private hands)

 When a national monument is designated around a private business in a liberal state, the ability of the President to alter its message is  at least partially circumscribed.

Trump Administration Removes Pride Flag From Stonewall National Monument  The enduring symbol of LGBTQ+ liberation has been taken down from the historic site.
By James Factora and Quispe López  February 10, 2026 

 

 

 "Manhattan borough president Brad Hoylman-Sigal told the New York Times that the directive to remove the Pride flag came from the Trump administration. The monument itself was designated in 2016 to honor the origin of Pride in the United States, and was also the first U.S. national monument dedicated to LGTBTQ+ rights.

"But like the 1969 rebellion that cemented Stonewall into history books, queer and trans people are not taking it without a fight. While the park and monument across from the original Stonewall Inn is now a federal park, the business itself is private property.

“Bad news for the Trump Administration: these colors don’t run,” Human Rights Campaign Press Secretary Brandon Wolf said in a statement. “The Stonewall Inn & Visitor’s Center is still privately owned, their flags are still flying high, and that community is just as queer as it was yesterday. While their policy agenda throws the country into chaos, the Trump administration is obsessed with trying to suffocate the joy and pride that Americans have for their communities.”

##########

N.Y.C. Officials Reinstate Pride Flag at Stonewall After Federal Removal   By Liam Stack and Olivia BensimonUpdated Feb. 13, 2026, 2:40 a.m. ET

"A group of New York elected officials gathered on Thursday to replace the Pride flag that was removed from the Stonewall National Monument after a directive from the Trump administration, mounting a defiant response to the government’s assault on diversity initiatives at a federal site honoring the L.G.B.T.Q. rights movement.

"The plan to re-raise the flag in the center of the small park outside the historic Stonewall Inn in Greenwich Village had been widely publicized on social media, and hundreds of spectators cheered as its rainbow colors made their way back up the flagpole under a cloudy winter sky."

Moral Economics: Book cover and jackets

 Getting a book out involves some tedium (e.g. trying to proofread the index) as well as many small excitements: here's the full book cover and jackets for Moral Economics:) 

 

 

 

Blurb I recently wrote (for a How-to on a familiar but surprising subject)

Like everything else in life, blurbs have editors, so not everything you write gets published.   

Some of the books I've read and blurbed might surprise you, such as this one (on a familiar subject, but a surprising one for a book):

You've Been Pooping All Wrong  by Dr. Trisha Pasricha (who I first encountered some years ago).

 Here's my blurb as it appears on the book's web page: 
“An entertaining and instructive book.”

  Here's the full blurb that I wrote

"Dr. Trisha Pasricha has written an entertaining and instructive book, in very plain language, about how our bodies turn inputs into outputs, along with tips on managing that. Along the way she writes equally clearly about the emerging, polysyllabic field of neurogastroenterology, which studies the lifelong, two-way conversation between brains and guts." 

 ##########

"My needs in life are simple, I want three things maybe four...a little love, just enough to eat, a warm place to sleep, and everything I write should be published"

Before I forget, coffee lowers risk of dementia

 Here's the good news, published yesterday in JAMA:

Coffee and Tea Intake, Dementia Risk, and Cognitive Function
Yu Zhang, MBBS1,2,3; Yuxi Liu, PhD2,3,4; Yanping Li, PhD1,2 et al,

JAMA, Published Online: February 9, 2026
doi: 10.1001/jama.2025.27259 

"Findings  In this prospective cohort study of 131 821 individuals from 2 cohorts with up to 43 years of follow-up, 11 033 dementia cases were documented. Higher caffeinated coffee intake was significantly associated with lower risk of dementia. Decaffeinated coffee intake was not significantly associated with dementia risk.

"Meaning  Higher caffeinated coffee intake was associated with more favorable cognitive outcomes. "

######

I figured I'd better pass along the news before I forget... 

Deadline tomorrow morning for Econometric Society conference on Economics and AI+ML

 Today's email announces a 9am Eastern time deadline tomorrow (that's 6am in California). And it looks like a fine conference.

FINAL CALL FOR PAPERS DEADLINE FEB. 10, 9:00 AM ET The 2026 ESIF Economics and AI+ML Meeting June 16 - 17, 2026 Reminder: authors can update their submitted papers until the submission period ends. The Econometric Society Interdisciplinary Frontiers (ESIF) conference on Economics and AI+ML will be hosted by Cornell University, in Ithaca NY, on June 16-17, 2026. The purpose of the meeting is to foster interaction of ideas and methodologies from the areas of Computer Science and Economics (broadly defined, but with emphasis on AI and ML). The conference will feature keynote lectures and parallel sessions, bringing together scholars from both fields. Important Dates Submissions open: November 3, 2025 Paper Submission Period: November 3, 2025 – February 10, 2026 Decision Notification Deadline: March 22, 2026 Registration Period (for presenters): March 22, 2026-April 5, 2026 Preliminary Program Announcement: April 26, 2026 Conference dates: June 16-17, 2026 Keynote Speakers David Blei, Columbia University Mingming Chen, Google Timothy Christensen, Yale University Annie Liang, Northwestern University Sendhil Mullainathan, MIT Aaron Roth, University of Pennsylvania Paper Submissions The deadline for submissions is February 10, 2026. Interested authors are encouraged to submit unpublished working papers or early drafts (more than 5 pages with results). Preference in the selection process will be given to complete papers. If you would like to submit a group of papers to be considered for the same session, please indicate the proposed session name in the comment section. While grouped submissions are welcome, please note that each paper will still be evaluated on its own merits. All papers and drafts must be submitted electronically in PDF format via the Oxford Abstracts submission platform.  Please use the appropriate submission link: Economist Track (Economics, Finance, Statistics, Marketing, Management) Computer Scientist track (Computer Science, Information Systems, and Operations Research/Operations Management) Submissions are open to all research that overlaps with both Economics and AI+ML. Each submission must select at least one content area from the drop-down menu. Note that the areas are partially overlapping; if in doubt, authors are advised to select those area(s) that best fit their paper. Also note that your submission details can be edited throughout the submission process up until the submission deadline of February 10, 2026. Please direct questions to esifaiml@gmail.com for more information, or visit the conference website. Program Committee Francesca Molinari and Eva Tardos (Cornell University; Co-Chairs) Nina Balcan, Carnegie Mellon University Sid Banerjee, Cornell University Dirk Bergemann, Yale University Martin Bichler, Technical University of Munich Larry Blume, Cornell University Emma Brunskill, Stanford University Flori Bunea, Cornell University Giacomo Calzolari, European Universitary Institute Denis Chetverikov, University of California Los Angeles Tim Christensen, Yale University Bruno Crepon, CREST Costis Daskalakis, MIT Sarah Dean, Cornell University Laura Doval, Columbia GSB David Easley, Cornell Maryam Farboodi, MIT Sloan Michal Feldman, Tel Aviv Christophe Gaillac, University of Geneva Avi Goldfarb, University of Toronto Rotman Jason Hartline, Northwestern Nicole Immorlica, Yale and MSR Jon Kleinberg, Cornell University Robert Kleinberg, Cornell University Anton Korinek, University of Virginia Elena Manresa, Princeton Mehryar Mohri, NYU, Google Xiaosheng Mu, Princeton University José Luis Montiel Olea, Cornell University Mallesh Pai, Rice David Parkes, Harvard David Pennock, Rutgers University Vianney Perchet, ENSAE Tuomas Sandholm, CMU Jon Schneider, Google Devavrat Shah, MIT Alex Slivkins, MSR Martin Spindler, Universität Hamburg Jörg Stoye, Cornell Vasilis Syrgkanis, Stanford Catherine Tucker, MIT Sloan