AI Is Already Disrupting Labor, and Recent Grads Could Be a ‘Lost Generation’ It’s a particularly bad time to be an unemployed 20-something.



“Artificial intelligence is going to replace literally half of all white-collar workers in the U.S.,” Ford CEO Jim Farley warned earlier this month. He’s not alone in sounding the alarm.

Perplexity CEO Aravind Srinivas recently told The Verge he expects AI to replace recruiters and executive assistants within six months. These warnings often come with short timelines, but some experts believe the disruption is not just coming—it has already started.

“The disruption of jobs is already underway, it’s expanding rapidly, and it will continue,” said John McCarthy, associate professor of global labor and work at Cornell University.

Signs of this shift are already visible. Anthropic’s newest AI assistant, released July 15, can perform nearly all the tasks expected of a finance intern at a typical Wall Street firm. Shopify CEO Tobias Lütke instructed hiring managers earlier this year that they must justify why a role can’t be done by AI before seeking approval to hire. Duolingo CEO Luis von Ahn sent a similar memo to his team.

According to McCarthy, this represents a fundamental threat to early white-collar careers. “We are facing a serious breakdown in the early stages of white-collar careers,” he told Gizmodo. “That’s really important because that’s where economic security begins—and right now, that foundation is being pulled away.”


A “Lost Generation” of Young Workers

The timing could not be worse for recent graduates.

The New York Fed reported in April that the labor market for 22- to 27-year-old college graduates “deteriorated noticeably in the first quarter of 2025,” with unemployment at its highest level since the pandemic. The gap between unemployment for young graduates and the overall workforce is now the widest since 1990.

Economic cooling and the end of a post-COVID hiring surge are part of the story—but AI is amplifying the problem. Generative AI excels at the very entry-level tasks new graduates are usually hired to perform.

“Evidence for AI’s negative impact on early careers is already strong,” McCarthy said. He worries that the current “generational squeeze” could harden into a permanent reconfiguration of career pathways.

Traditionally, entry-level white-collar work has functioned as an on-ramp to long-term career stability. With fewer of these roles available, young workers are increasingly relying on elite internships, personal connections, and networking—dynamics that will likely widen inequality.

“There is a real fear that an entire cohort—those graduating during the early AI transition—may become a lost generation,” McCarthy warned. “Unless policy, education, and hiring norms adjust, and I’m not tremendously optimistic about those adjustments happening at the necessary scale.”


Is AI Really to Blame?

Not everyone believes AI is the core problem.

Robert Seamans, a professor of management at New York University, argues that fears of an AI-driven labor crisis may be overstated. “We’re actually seeing relatively low rates of AI adoption across the corporate sector,” Seamans told Gizmodo.

A recent Federal Reserve paper supports his view. While tech, finance, and research-heavy industries are experimenting with AI, most companies—particularly smaller ones—have yet to integrate these tools into daily operations. Even in larger firms, implementation hurdles remain.

“It’s much harder to implement AI in a firm than people realize,” Seamans explained. “Firms often lack the in-house talent to train, operate, and oversee the AI they want to use. Until you have that expertise, it’s hard to rely heavily on the technology.”

Some companies, he added, might be using AI as a convenient scapegoat for hiring freezes or headcount reductions. “It’s much easier to blame AI than tariffs or economic uncertainty,” Seamans said.

Still, he acknowledges that AI likely plays at least some role in current labor market trends. He believes a “fully funded U.S. statistical agency” like the Census Bureau or Bureau of Labor Statistics should track AI deployment in real time to separate hype from reality.


The Road Ahead

AI is not going away, and its influence on the workforce will only grow. McCarthy predicts that the technology will not necessarily trigger “wholesale termination” of jobs, but it will lead to a profound restructuring of work.

“Human work is shifting and it will continue to shift,” McCarthy said. “Roles that require judgment, ethics, creativity, or contextual understanding will endure.”

This shift will pressure both higher education and K–12 schools to rethink how they prepare students. At Cornell, McCarthy is already integrating AI-assisted workflows into his classes, alongside teaching adaptable, generalizable skills.

Policy is the other critical piece. “These changes are happening very fast and with greater potential to impact jobs at scale than any point in history,” McCarthy said. He argues that government, schools, and private companies must collaborate on how to respond.

For workers navigating this transition, the best strategy is to become fluent in AI tools, stay adaptable, and be ready to pivot.

“I don’t say this lightly,” McCarthy said. “I have a 7-year-old, and I worry very much about what the future of work will look like for him.”


In early March, Volkan Çinar, a postdoctoral researcher in chemistry at MIT, received an unexpected email that would shift the trajectory of his career. The message invited him to join a new fellowship program training AI models—work closely tied to his expertise in carbon-carbon bond formation in graphene. With academic positions increasingly scarce and fiercely competitive, Çinar began questioning whether a future in academia still made sense. The offer came at just the right time.

The program, called **MOVE**, was launched on June 10 by **Handshake**, a $3.5 billion-valued platform originally built to connect students with employers. Since its founding in 2014, Handshake has amassed a vast network of university students and alumni across the U.S. Now, it’s leveraging that network to bridge a growing gap in the AI economy: the need for domain experts to train generative AI models.

Garret Lord, Handshake’s founder, conceived the idea after conversations with researchers from leading AI labs. He learned that many PhDs and master’s students were being recruited informally to help fine-tune AI systems—but often faced inconsistent pay and poor training. Recognizing Handshake’s unique position, Lord saw an opportunity: *What if we could professionalize this work?*

“We wanted to ensure that as the world of work evolves into an AI-driven economy,” says Christine Cruzvergara, Handshake’s Chief Education Strategy Officer, “our students are equipped with the skills and experiences needed to thrive.”

 How the MOVE Fellowship Works

MOVE connects graduate students and postdocs like Çinar with short-term, high-impact projects at generative AI labs. Fellows typically commit 10 to 20 hours per week and earn between **$40 and $130 per hour**, depending on their expertise.

Before starting, participants complete a 2–5 hour foundational training module covering how large language models function and best practices for prompt engineering. Some projects require additional onboarding, but all emphasize precision, clarity, and critical thinking.

Due to strict NDAs, fellows can’t disclose specifics about their AI partners or tasks. But Rachel Mitchell, a doctoral candidate in education at the University of Miami, described the general workflow.

She works in collaborative “pods” with other subject-matter experts to design prompts or real-world scenarios for AI models to respond to. After analyzing the output, the pod evaluates what the model got right—and where it failed. Was the error due to flawed reasoning, or was the prompt itself ambiguous? They refine the input, retest, and provide structured feedback to the AI lab, which uses it to improve model performance.

“I want to make sure we’re actually advancing the AI’s knowledge,” Mitchell says, “not just exposing weaknesses caused by unclear instructions.”

Her goal is to help build tools that educators can trust. Many teachers today, stretched thin and overwhelmed, are turning to AI to generate lesson plans. But without proper grounding in educational research, these outputs often lack measurable learning objectives or evidence-based strategies. Mitchell believes that if AI is trained by experts who understand pedagogy, it can become a powerful ally in classrooms.

Similarly, Çinar sees AI as a tool to accelerate scientific discovery. Literature reviews in chemistry can take weeks; AI could reduce that to hours—if it’s properly guided. “This experience,” he says, “is teaching me how to communicate complex ideas to machines, which will be essential no matter where I end up.”

Beyond technical skills, Çinar values the soft benefits: networking with peers across disciplines, practicing project delivery under deadlines, and gaining exposure to corporate workflows. “Learning how to operate in a professional environment,” he notes, “is something you don’t always get in the lab.”

A Human-Centric Approach to Talent Matching

Unlike platforms like Surge AI, which rely on algorithms to assign tasks, **Handshake uses human reviewers** to match fellows with projects. Applications are evaluated based on résumés, interviews, and performance in the training module. Domain experts review candidates’ qualifications and often serve as project leads, offering mentorship and feedback throughout.

This curated approach ensures only true specialists are selected—no generalists. “We’re looking for people who can *challenge* AI,” Cruzvergara emphasizes. “Not just follow instructions, but think deeply about how models reason, where they fail, and why.”

Applicants stand out by showcasing advanced research, publications, patents, or teaching experience. Demonstrated problem-solving ability, experience designing assessments or rubrics, and a history of collaboration—especially in interdisciplinary teams—are key differentiators.

Democratizing Access to the AI Economy

As demand for AI skills skyrockets—from 3,000 job postings in 2010 to over 80,000 in 2024, according to Brookings—there remains a stark gap in access to training. A 2024 Pew survey found that **63% of Americans have never used AI at work**, and only **24% have received formal AI training**.

Handshake aims to close that gap—especially for students in fields like the humanities, where tech opportunities have historically been limited. Emory University, one of the first institutions to partner with MOVE, views the fellowship as a vital step toward preparing students for the future of work.

After a presentation from Handshake in March, Emory’s leadership signed on immediately. “We saw this as a chance to give our students hands-on experience with AI in meaningful, domain-specific ways,” said Branden Grimmett, Associate Dean at Emory College and Vice Provost for Career and Professional Development.


Within weeks, **110 Emory students applied**—a level of interest far exceeding most career programs. Nationally, Handshake reports a **25% acceptance rate**, though the program is still too new for any fellow to have yet converted their experience into a full-time role.

Still, the long-term vision is clear.

Lord envisions a future where MOVE fellows earn verifiable **badges** for their contributions, schools compete on public **leaderboards**, and top performers are fast-tracked into AI careers. The company is also exploring expanding eligibility to young professionals beyond academia.

The Future Is AI-Literate

Cruzvergara puts it bluntly: “I don’t believe AI will take your job. But I do believe someone *using AI effectively* will.”

For students and early-career researchers navigating an uncertain job market, programs like MOVE offer more than just income—they provide **relevance, adaptability, and agency** in an era of rapid technological change.

Whether Çinar returns to research or transitions into industry, one thing is certain: his ability to shape AI from the inside will make him a more competitive candidate. And for thousands of others watching from labs, libraries, and lecture halls, MOVE may represent not just a side gig—but a lifeline into the next economy.


**Paid subscribers will learn:**

- What to expect from the MOVE fellowship  

- The program’s acceptance rate and pay structure  

- How to position yourself as a competitive applicant  

- Why AI labs are turning to expert networks instead of crowdsourced labor  


Handshake isn’t just connecting talent to jobs anymore. It’s helping define what the next generation of work will look like—one expert, one prompt, one AI model at a time.

Post a Comment

Previous Post Next Post