A wave of dropouts is sweeping through America's elite universities, driven by fears and opportunities surrounding artificial general intelligence (AGI). Students from Harvard and MIT are abandoning their studies, split between "accelerationists" seeking to capitalize on AI breakthroughs and "doomers" joining AI safety organizations to prevent potential human extinction.

**Elite University Dropout Wave Begins at Harvard and MIT**

In the hallowed halls of Harvard University and MIT, a silent "dropout wave" is spreading among students once destined for greatness in technology and academia. These elite students are voluntarily withdrawing from their studies, driven not by confusion but by fear—deep anxiety about the imminent arrival of AGI.

This phenomenon has intensified in 2025. Students believe that once AGI is truly achieved, it will comprehensively surpass humans in almost all fields and may even bring catastrophic consequences. "If AGI is coming and will lead to human extinction, what meaning does my degree have?" they ask.

According to reports, dozens of students from Harvard and MIT have chosen to drop out this year alone. They are directing their energy toward AI safety startups and research labs, attempting to promote AI "alignment" to ensure AI follows human values.

**Dropout Students: "I Worry I Might Not Live to Graduate"**

Alice Blair entered MIT in 2023 and joined an AI ethics group, seeking balance between technology and conscience. However, a year later, she chose permanent leave to join an AI Safety Center as a technical writer. She believes: "In most cases, the way we strive to achieve AGI will lead to human extinction."

Alice's concerns are not unfounded. Many researchers are exploring these issues. In 2017, Professor Stuart Russell emphasized AI safety's importance with a vivid analogy: if we cannot control more intelligent AI systems, it's like building an atomic bomb without being able to control the explosion.

Last year, "godfather of deep learning" Geoffrey Hinton warned that AI could potentially lead to human extinction, with AI development progressing faster than he expected. In 2024, a U.S. State Department report indicated that uncontrolled AI development indeed poses "extinction-level" risks.

Numerous tech leaders have weighed in on AI development timelines. Sam Altman predicts AGI will appear before 2029, while Demis Hassabis considers 5-10 years a reasonable timeline. AGI may be accelerating toward us.

Alice is not the only fearful student. Harvard sophomore Adam Kaufman chose to take leave and join Redwood Research, which studies "deceptive AI systems." He stated: "I'm really worried about AI risks, and I think the most important thing is to mitigate this risk. This field is so important that it's worth my full-time commitment."

Adam's brother, roommate, and girlfriend have all left Harvard and now work at OpenAI. This "collective dropout" behavior has evolved into a group phenomenon of "AI anxiety."

**AI Anxiety: America's Last Generation of Graduates?**

A Harvard survey shows that more than half of students worry that AGI will affect their employment prospects. "If your career will be replaced by AI before 2030, then every year you spend in college is actually compressing your brief professional career."

From GPT-5 to autonomous AI systems, technology's rapid advancement convinces many students that university curricula can never keep pace with AI development. Harvard's early-year "AI Anxiety" research revealed that undergraduates worry AI will render human skills obsolete, especially in creative fields.

Dario Amodei publicly stated that AI will replace half of entry-level white-collar positions. Unemployment rates could soar to 20% in the coming years. Media predictions suggest America's next economic crisis could bring massive layoffs.

In past economic cycles, whenever recession occurred, companies often accelerated automation processes, reducing long-term dependence on human labor. For example, U.S. manufacturing employment decreased by 26% from 2000 to 2019, largely due to automation-driven efficiency improvements.

J.P. Morgan senior U.S. economist Murat Tasci wrote in a report: "We believe that in the next economic recession, AI tools and applications will be rapidly and widely adopted in workplaces, potentially leading to the replacement of many jobs primarily involving non-repetitive intellectual tasks."

Federal Reserve Governor Lisa Cook stated last month: "AI has the potential to reshape our labor markets, which could also affect our understanding of 'full employment' or our assessment of the natural unemployment rate. Like many technological breakthroughs, some jobs may be replaced."

**AI's "Tulip Mania"?**

Not everyone subscribes to this narrative. Meta's Chief AI Scientist Yann LeCun publicly refutes these claims, stating that "existing large models have less learning capability than a cat." AGI safety advocate Gary Marcus even bet against Elon Musk, believing "we won't see superhuman AGI by the end of 2025."

However, for today's young people, their college careers are dominated by one question: Will the future world still need humans?

On social media, posts under the AI dropout hashtag go viral, with one highly-liked tweet reading: "The smartest generation is leaving school to guide AGI's birth positively."

Skeptical voices exist too. Harvard and MIT's joint research published last month in "Algorithmic Bridge" pointed out that current AI models are far from achieving true scientific discovery capabilities, suggesting panic may be exaggerated.

For dropouts like Alice Blair, the risks cannot be ignored. She asserts: "Preventing AI from turning against humanity is our generation's defining challenge."

Some voices try to hit the pause button. Alice Blair herself doesn't think everyone should drop out: only those resilient enough, whom college has already taught to live independently, are suitable for early departure.

Y Combinator co-founder and Sam Altman's former boss Paul Graham posted on X: "Don't drop out to start a company. Opportunities will come again, but college time won't return." In February this year, he encountered an 18-year-old entrepreneur duo who received $75 investment for a paper and decided to drop out to start a business. He believed this subtle encouragement of student dropouts crossed a line.

Harvard has even urgently launched AI ethics courses, while MIT removed a paper promoting AI productivity improvements. Universities are trying to retain students through courses and persuasion.

But this generation of students understands the reasoning—they just feel the world is changing too fast to wait any longer. A prominent economic commentator wrote on social media: "In Gen Z's eyes, degrees have become souvenirs from before ChatGPT appeared."