Computing Is Indeed a Discipline in Crisis

In July 2024, I raised the question of whether computing is a discipline in crisis. The trigger for that was a January 2024 survey of thousands of artificial intelligence (AI) authors on the future of AI, whose main message was that the AI community has deep concerns about the direction of the field but feels unable to influence that direction, which is mostly driven by the Silicon Valley money machine. That concern has only been sharpened since then. In fact, Eric Schmidt, a Silicon Valley insider, recently penned an opinion article with Selina Xu in The New York Times titled “Silicon Valley Is Drifting Out of Touch with the Rest of America,” stating: “This frenzy gives us pause.”

While Silicon Valley is investing tens of billions of dollars chasing the artificial general intelligence dream, academic computing research in the U.S. is facing a severe drought. As I pointed out in April of this year, U.S. academic research has already been dealing with an intensifying competition for flat funding over the past 15 years, and while budget predictions for the next fiscal year are still uncertain, it seems clearer that the federal funding budget for computing research will shrink considerably.

The picture that emerged at the Computing Research Association Summit last July is one of several converging storms. For example, in view of the growing gap between increasingly tight federal funding for academic computing research and the widely generous industrial funding, many AI researchers are questioning the viability of the academic path. “AI researchers are negotiating $250 million pay packages,” reported The New York Times recently.

Beyond the funding issue, there is the question of who will carry out the research in academia. Academic research in engineering and science is carried out by doctoral and postdoctoral students under the supervision of faculty members. About 90% of U.S. graduate programs in computer science and engineering have a majority of international students, but the attractiveness of the U.S. as a place to pursue advanced study has been diminishing for years. NAFSA, the Association of International Educators, predicts a 15% drop in overall international enrollments (graduate and undergraduate) this fall.

The unrelenting growth of academic computing departments over the past 15 years has been driven by the attractiveness of computing as a career. I still remember that computer-science students would tell me they chose to pursue computer science because they enjoyed programming in high school. That later changed to “computer science offers good jobs,” which later changed to “computer science offers good money.” That was then. In early August 2025, an article in The New York Times with the sensational title “Goodbye, $165,000 Tech Jobs. Student Coders Seek Work at Chipotle” described that, “As companies like Amazon and Microsoft lay off workers and embrace AI coding tools, computer science graduates say they’re struggling to land tech jobs.” In view of such national news, it is a certainty that computer-science enrollments, which have been rising for years, will decline significantly.

At the same time, a critical MIT report revealed that 95% of enterprise generative-AI programs are failing, triggering speculations about the next “AI Winter.” In fact, the phrase “irrational exuberance,” used by Alan Greenspan in a 1996 speech given during the dot.com bubble of the 1990s, seems an apt metaphor to describe today’s dot.AI frenzy.

And yet the bubble extends to more than just stock-market valuations. A recent email from the Association for the Advancement of AI (AAAI) informed the community that “AAAI-26 received almost 29,000 submissions to the Main Technical Track. After removing papers that were not fully compliant with submission policies, we still have roughly 23,000 papers under review—nearly twice the number of papers reviewed by AAAI-25. AAAI-26 saw a much higher-than-expected number of submissions from China—almost 20,000 of the 29,000 total—a welcome sign of global engagement.” On one hand, we could rejoice in the blossoming of globalized AI research. On the other hand, our idea of a research conference as a venue for scientific exchange grounded in a well-curated set of accepted papers clearly does not scale to 23,000 bona fide submissions. I suspect this is the bitter lesson of this decade.

Computing has always been a seesawing discipline. We had the AI winters of the 1970s and the 1990s, and we had the dot.com crash and the ensuing “Image Crisis.” It seems another crisis is upon us. Will we ever learn? It is time for some serious thinking!