This summer, fetching lattes for executives is out. Teaching bosses how to leverage artificial intelligence is in.
When Jasreen Toor interned at the agentic human resources company Eightfold AI this summer, her days were not filled with answering phones or taking meeting notes. Instead, she spent the majority of her time building and monitoring a Claude AI agent designed to scour company data and generate first drafts of marketing materials. While she had used AI to brainstorm and research in previous internships, at Eightfold, AI *was* the work.
“A lot of what I was doing was taking my own initiative and building on my own to see how we could solve some of the problems that product marketing was facing,” says Toor, a rising junior studying economics and data science at the University of California, Berkeley.
The Internship Paradox
For years, experts have warned that AI could collapse the white-collar career ladder, devouring entry-level jobs and leaving recent graduates without foundational rungs to climb. The data reflects this anxiety: recent college graduates currently face a 5.7% unemployment rate, nearly double the 2.9% rate for all college graduates in the workforce, according to the Federal Reserve Bank of New York.
Yet, paradoxically, the National Association of Colleges and Employers (NACE) predicts the number of internships will grow by 4% year-over-year. Today’s interns have become the ultimate testing ground for how AI will reshape early-career work. As chatbots and autonomous agents absorb traditional grunt work—note-taking, data entry, and slide deck formatting—some companies are trusting interns with unprecedented responsibility, tasking them with designing the very AI-driven workflows that replace those old duties.
However, this shift may be fleeting. Every tedious process automated this summer eliminates a tried-and-true training task for next year’s cohort. Every new workflow or AI agent developed by a 2026 intern will eventually fall to someone else to maintain—or quietly abandon.
From "AI Slop Janitors" to Strategic Architects
The unlucky members of the Class of 2026 may have spent their summer as "AI slop janitors," endlessly sorting through AI-generated code or documents to catch basic, rote errors—the 2020s equivalent of filing paperwork. But the fortunate ones have hit the sweet spot, playing a tangible role in shaping how their companies will operate before the technology fully upends junior-level work.
“The best companies for interns right now are the ones that allow young workers to use AI to transform the role itself,” says Shanda Mints, Vice President of AI Strategy and Transformation at Korn Ferry. “But that’s only going to be around for the next couple of years. Once that role becomes transformed, what are the new interns going to be doing a few years from now? We’re seeing that they’re likely going to just jump right into what historically were the entry-level roles.”
The career ladder may not be collapsing entirely, but it is undeniably shrinking as the responsibilities placed on young workers intensify.
The New Career Catch-22
Hiring managers are increasingly bypassing traditional résumé screening in favor of demonstrable, hands-on skills, exacerbating the classic career Catch-22: you need experience to get experience. According to a recent NACE survey of employers, the proportion of entry-level jobs requiring AI skills surged from 13.5% to 35% between last fall and this spring.
Consequently, an internship where a student builds an AI solution to a team-wide efficiency problem has become a golden ticket in a constricted hiring market. Companies are taking notice:
* **Liberty Mutual** offered full-time jobs to 79% of its rising senior interns last year. This summer, as AI absorbed traditional tasks like data entry and document review, interns in actuarial and analytics roles used AI to analyze data, visualize trends, and edit code. The time saved allowed them to craft presentations for senior management, requiring a deep understanding of how their findings impacted the broader business model. “That was the work that actually built discipline, but it didn’t always build judgment,” says Maura Quinn, VP of Early Talent Acquisition and Engagement Programs at Liberty Mutual. “That’s where we’re seeing the big shift: AI tools are handling a lot of that routine processing.”
* **Duolingo** is similarly emphasizing AI fluency. “Interns may engage in casual ‘vibe coding,’ but they do so alongside other team members, rather than developing siloed solutions to automate their own work,” says Debica Bhattacharya, Head of People. Duolingo is highly motivated to prepare these graduates for the workforce, having extended full-time offers to 80% of its interns last year.
Building the Future, One Agent at a Time
At MiQ, a programmatic media company, senior account manager Lindsay Foxen oversaw fellows this summer as the company integrated more AI into its workflows. The fellows built custom agents: one to compare internal documents for consistency, and another to cross-reference data input in Salesforce with revised data, highlighting discrepancies and eliminating the need for line-by-line manual checks.
The fellows were given the time to observe MiQ’s operations and then implement “what they would love to see from an AI-automated tool,” Foxen notes. The solutions they created prove “that they understand AI, they understand MiQ, and they understand the future of what’s being implemented moving forward.”
This demand for AI-ready talent is surging on campus, too. At Northeastern University, which integrates six-month work periods (co-ops) into its academic schedule, the number of students placed into co-ops requiring AI skills has doubled year-over-year to 1,500, according to Chancellor Ken Henderson. “What each student on a co-op can actually produce of value is higher as a function of the technology,” he explains. “They can demonstrate a work product that’s useful for their future careers, doing something tangible and valuable during that time.”
Matt Mizra, a rising senior studying computer science and business at Northeastern with internships at Microsoft and DraftKings, says he has been encouraged to leverage AI as much as possible this year. Rather than churning out basic software engineering tickets, he spends the bulk of his time planning architecture and prompting Claude to write the code. While he misses the act of writing code himself, the scope of his work has elevated significantly. “My projects are very big, and they have good business value,” Mizra says. “It’s more about thinking about how to solve the bigger picture.”
Yet, Mizra remains wary of the early-career job market. Tech companies are actively battling AI cheating in engineering interviews, and the bar for technical assessments keeps rising. “It’s definitely making it harder, because everyone is proficient at code now.”
The Looming Crisis of Judgment
As AI seeps deeper into both the college experience and the daily workday, fears are mounting that overreliance will erode critical thinking. Colleges lack consensus on whether—or how—to enforce analog learning, while companies grapple with the best ways to audit and review their workers’ AI usage. Early research continues to raise valid questions about the long-term effects of AI on human creative thinking and reasoning skills.
Caught in this uncertainty, young workers eagerly seek the skills and security most likely to lead to full-time employment. However, the demands of these jobs may be escalating more quickly than young workers can develop confident, independent judgment.
Ultimately, the best training ground still comes from doing the work. Managers who want to train future talent on how to use AI must ask themselves a critical question, according to Korn Ferry’s Mints: “Is that task going to help elevate this person’s judgment? Is that experience enough to replace the experience they would have been getting if they had whatever their internship was to begin with?”
Thanks to the pioneering, AI-driven work of the 2026 intern class, next year’s summer interns will walk into workplaces that look vastly different. The challenge for employers will be ensuring those workplaces still know how to build a leader.