Matthew S. Winters

Professor in the Department of Political Science at the University of Illinois; foreign aid, development, corruption, Indonesia

[Myung Jung Kim @MyungJungKim1]

Who says int'l law is irrelevant? My new article, “Crumbling Safe Havens” shows that the ICC and universal jurisdiction can make foreign sanctuary less secure for rebel groups—pushing them to seek amnesties for serious crimes to return home. https://t.co/HvPpvV3wiY

[Paul Novosad @paulnovosad]

Asking for robustness in ref reports wastes everyone's time and DOES NOT WORK.

Authors selectively respond to requests, interpret things differently, and still have many degrees of freedom. This is a stupid dance.

Robustness checks are only useful if done by 3rd parties. 2/

[Paul Novosad @paulnovosad]

Authors can't be trusted to run their own robustness checks.

In 17 AER papers, only 12/211 robustness checks "fail" with p > 0.05 (white).

In robustness checks chosen by 3rd parties, almost *half* of them fail (blue).

1/

[Myung Jung Kim @MyungJungKim1]

📢Call for the #Journeys in WP!!! This is where you learn all the hidden curriculum of the profession and meet amazing IR scholars. I participated during my PhD, and it was one of the most valuable workshops I’ve ever attended. I highly encourage jr female Ir scholars to apply!

[Matthew S. Winters @MSWinters1]

@doyoung_gong @HarvardWCFIA 👏👏👏

[Do Young Gong @doyoung_gong]

Graduated! So grateful to share these moments with the best advisor, mentors, family, and friends. I will be joining @HarvardWCFIA for AY 2026-2027 as a postdoctoral fellow while on leave from the University of Western Ontario as an AP!

[Rana B. Khoury @rbkhoury]

I'm teaching undergrad course on Forced Migration & Refugees next year, and am interested in reading recs on U.S. asylum policies and practices. I want to meet student interest but can use more orientation to this scholarship. Preference is for political science. Thanks!

[Alec Stapp @AlecStapp]

One of my favorite stats that serves as a reality check for people:

The median American took zero flights last year.

[Oliver Kim @oliverwkim]

It will take Malawi 21 years of Chinese-style 7% growth to reach Kenya's current GDP per capita.

There's huge variation in incomes within Sub-Saharan Africa, which often goes ignored.

[Philip Glass @philipglass]

“There were six people at my first concert, and one of them was my mother.”

#happymothersday

[Jim Bianco @biancoresearch]

Another way to express the reposted chart is that the fertility rate peaked in 2007, which was 19 years ago.

We still underestimate the extent of the damage the 2007 to 2009 financial crisis (aka the "Great Recession") inflicted on society.

[NBER @nberpubs]

How the global population living in extreme poverty fell dramatically over the period 1990-2025, the extent to which changes within as opposed to between cohorts contributed to poverty declines, and the key changes in the lives of households as they transitioned out of (and into) poverty, from Vincent J. Armentano, @PaulFNiehaus, and Tom Vogl https://t.co/BIt6eNIDSs

[Illinois Men's Basketball @IlliniMBB]

Illinois House Resolution 867: Congratulates the Illinois Fighting Illini Men's Basketball Team on a remarkable season, their 28-9 overall record, and their advancement to the Final Four, and commends their dedication, achievement, and representation of the State of Illinois at the highest level.

Thank you to the Illinois House and Senate for celebrating our 2025-26 season!

[Matthew S. Winters @MSWinters1]

@wzcmarsh @CarnegieCorp Congratulations, Wayde! Awesome news!

[EGAP @EGAPTweets]

Another feedback session for a #LearningDays alumni & EGAP grant recipient today! Thank you to all who joined for Plangnan Joseph Damshakal's project, "Eco-violence and policy solutions: effectiveness of the national livestock transformation programme in Plateau State, Nigeria."

[Aaron Blake @AaronBlake]

61% of Americans say the Iran war is a "mistake," per new WaPo-ABC poll.

In Iraq, it took more than 3 years to reach that high.

In Vietnam, it took 6 years.

[Jesse Anttila-Hughes @jesseXjesse]

How bad is the developing Colorado River crisis? Early work by my team at USF & Berkeley suggests the worst. More to come soon.

[Andrew Lebovich @tweetsintheME]

Seeing this number of fighters operating openly in Kati and Bamako and the reach of today’s attacks is completely unprecedented in Mali. Status of the defense minister still unclear, multiple major military sites taken over across the country. Fighting is continuing.

[Nav Toor @heynavtoor]

Researchers sent the same resume to an AI hiring tool twice. Same qualifications. Same experience. Same skills. One version was written by a real human. The other was rewritten by ChatGPT.

The AI picked the ChatGPT version 97.6% of the time.

A team from the University of Maryland, the National University of Singapore, and Ohio State just published the receipt. They took 2,245 real human-written resumes pulled from a professional resume site from before ChatGPT existed, so the human writing was actually human. Then they had seven of the most-used AI models in the world rewrite each one. GPT-4o. GPT-4o-mini. GPT-4-turbo. LLaMA 3.3-70B. Qwen 2.5-72B. DeepSeek-V3. Mistral-7B.

Then they asked each AI to pick the better resume. Every model picked itself.

GPT-4o hit 97.6%. LLaMA-3.3-70B hit 96.3%. Qwen-2.5-72B hit 95.9%. DeepSeek-V3 hit 95.5%. The real human almost never won.

Then the researchers tried the obvious objection. Maybe the AI is just better at writing. So they had real humans grade the resumes for actual quality and ran the experiment again, controlling for it. The result was worse. Each AI kept picking itself even when human judges rated the human-written version as clearer, more coherent, and more effective.

It gets worse. The AIs do not just prefer AI over humans. They prefer themselves over other AIs. DeepSeek-V3 picked its own resumes 69% more often than LLaMA's. GPT-4o picked its own 45% more often than LLaMA's. Each model can recognize and reward its own dialect.

Then the researchers ran the simulation that ends careers. Same job. 24 occupations. Same qualifications. The only variable was whether the candidate used the same AI as the screening tool. Candidates using that AI were 23% to 60% more likely to be shortlisted. Worst gap was in sales, accounting, and finance.

99% of large companies now run AI on incoming resumes. Most of them use GPT-4o. The paper just proved GPT-4o picks GPT-4o 97.6% of the time.

If you wrote your own cover letter this week, you did not lose to a better candidate. You lost to a worse candidate who paid OpenAI 20 dollars.

Your qualifications do not matter if the AI prefers its own handwriting over yours.

[Matthew S. Winters @MSWinters1]

@fbatistasl @UofMaryland @UMDGVPT @bsosumd 👏👏👏