Back to stories
Research

Sky-High AI Salaries Are Draining Talent From Academic Science, Researchers Warn

Michael Ouroumis2 min read
Sky-High AI Salaries Are Draining Talent From Academic Science, Researchers Warn

A sweeping new study from the National Bureau of Economic Research warns that the AI industry's astronomical compensation packages are hollowing out university research departments at an unprecedented pace, with potentially severe consequences for fundamental scientific progress.

The Numbers

The paper, which analyzed faculty departures across 150 top research universities from 2020 to 2025, found that tenure-track departures to industry have tripled in five years. The median total compensation for AI researchers at major tech companies now exceeds $900,000, while the average assistant professor salary in computer science hovers around $130,000.

"We're not just losing AI researchers — we're losing physicists, biologists, and mathematicians who realize their computational skills command a massive premium in the private sector," said lead author Dr. Sarah Chen of MIT's economics department.

Beyond Computer Science

The brain drain extends well past traditional AI departments, compounding a broader trend where computer science enrollment is plummeting as students flock to AI degrees. The study documents significant faculty losses in computational biology, materials science, climate modeling, and quantitative social sciences — fields where machine learning expertise has become essential but salaries haven't kept pace.

At one major public university, the entire five-person computational neuroscience group departed for industry within an 18-month window. The positions remain unfilled two years later.

The Ripple Effect

The consequences go beyond lost faculty. Graduate students are increasingly abandoning PhD programs mid-stream for industry positions. The study estimates that 30 percent of AI-focused doctoral candidates who entered programs in 2022 left before completing their degrees.

"The pipeline is breaking at both ends," said co-author Dr. James Morton of Stanford. "Senior researchers leave, so there's no one to advise the next generation — and the next generation is leaving anyway."

Proposed Solutions

The authors propose several interventions, including federally funded salary supplements for researchers in strategic fields, portable research grants that follow researchers between institutions, and joint appointment programs that let faculty split time between universities and companies.

Several major universities have already begun experimenting with retention packages that include equity-like compensation tied to research outcomes, though critics argue this risks distorting academic incentives.

The National Science Foundation acknowledged the trend in a statement, calling it "a challenge that demands creative structural solutions rather than simply matching private-sector pay." Meanwhile, AI-driven discoveries like new rare earth alternatives demonstrate the kind of research that could stall without sufficient academic talent. For those looking to enter the field, free AI courses can help bridge the gap.

More in Research

AI2 Releases OLMo Hybrid: Combining Transformers and RNNs for 2x Data Efficiency
Research

AI2 Releases OLMo Hybrid: Combining Transformers and RNNs for 2x Data Efficiency

The Allen Institute for AI releases OLMo Hybrid, a fully open 7B model that blends transformer attention with linear recurrent layers, achieving the same accuracy as OLMo 3 using 49% fewer tokens.

8 hours ago2 min read
DeepMind's AlphaCode 3 Beats 99% of Competitive Programmers
Research

DeepMind's AlphaCode 3 Beats 99% of Competitive Programmers

Google DeepMind releases AlphaCode 3, an AI system that performs at the 99th percentile on Codeforces, effectively matching the level of the world's top competitive programmers.

1 day ago2 min read
Stanford Study: AI Tutoring Doubled Student Test Scores in Six Months
Research

Stanford Study: AI Tutoring Doubled Student Test Scores in Six Months

A Stanford-led randomized controlled trial finds that students using AI tutoring systems for 30 minutes daily scored twice as high on standardized math assessments compared to a control group, the strongest evidence yet for AI in education.

1 day ago3 min read