Emma Kanjorski doesn’t consider herself an advanced AI user — at least compared to her fellow 2026 graduates. For most of her time at the University of Vermont she avoided ChatGPT because she did not want to cut into it.
Ultimately, however, he discovered how AI could help him parse dense financial reports and process data. By senior year, the business major was advising younger classmates on how to use AI to check their case study work and showing a professor how to induce a “sanity check,” or getting an AI to critique one’s own output.
Now Kanjorski sees AI as a potential edge when she starts this summer as a financial analyst at an insurer in Burlington, a position she feels lucky to find after applying to about 40 jobs. “I would like to be that person who can help other people understand it better or figure out how it fits into their day-to-day,” she said.
Here comes the Class of AI, the most AI-native group of graduates to enter the workforce — a group employers are already trying to figure out what to do with. He started college just a few months before ChatGPT came into the world. They’re leaving as AI is increasingly shaking up entry-level jobs that were once considered solid career launchpads.
Compared to his predecessors, he has an innate versatility with rapidly evolving technology and little respect for the notion that he has to repeatedly pay his dues with grunt work. In a recent Gallup-Lumina Foundation survey of nearly 6,000 Americans, 22% of 18- to 24-year-olds with a two- or four-year degree said they feel “very prepared” to compete in an AI-shaped job market, more than any other age group.
“We’re asking for the entire workforce to be reskilled, but in reality, only new graduates have the tools to get that experience,” said Allison Srivastava, education and labor economist at Niche, a college ranking and review site.
SharkNinja CEO Mark Barrocas invited two dozen students to a two-day AI hackathon in April to create tools, including one that crunches market-trend data to help identify new potential product lines. “What we’re finding is that these young people have a better opportunity to make an impact today than they’ve ever had before,” he said. “The AI skills they are bringing are more advanced than someone with 20 years of experience.”
The device maker is hiring about 200 “AI-forward” graduates and trainees this year, including about 10 from the hackathon. Companies like IBM, Salesforce and MetLife say they are also increasing their hiring of fresh graduates to leverage their AI-native skills.
Yet elsewhere, graduate hires have been the first to suffer due to corporate cost cutting in the name of AI and its ability to perform entry-level tasks such as coding and putting together slide decks. Unemployment among college graduates aged 22 to 27 was 5.6% in March, one of the highest rates since the early 2013 pandemic.
A survey of nearly 1,500 employers published last week by the Strada Education Foundation shows an ambivalence about hiring college graduates: Three times as many of the companies investing in AI said they expect entry-level hiring this year to increase than decrease. Nevertheless, the share in junior appointments is projected to be cut from 13% to 17% in 2025. The survey did not ask companies to quantify their recruitment plans.
This is a big reason why the Class of 2026 has a conflicted relationship with AI. Commencement speakers, including former Google CEO Eric Schmidt, faced criticism when they invoked AI at graduation ceremonies this month. Even after her job ends, Kanjorski said she’s worried about the future, and her mother frequently emails her articles warning of an AI job apocalypse. “I’m like, ‘Mom, I’m already scared,'” she said.
Leila Hernandez, a new graduate of San Diego State University who is still looking for a job as an accountant, put her feelings about AI more bluntly: “I wish it wasn’t here.” If she doesn’t get a position soon, she may look for work in some other field, she said.
New graduates are also concerned about other aspects of AI. In a December RAND survey, nearly two-thirds of college students who use AI for homework agreed that the technology was harming critical thinking skills.
“There are two parts of me that war against each other with AI,” said Naomi Sato, who graduated last week from Chapman University in Orange, California, with a degree in graphic design. The first time she used ChatGPT her freshman year — for romance-novel recommendations — she was disappointed: Everything it recommended was obvious, she said. Although her professors warned against using it to cheat on assignments, she was not tempted. People got six thumbs up from the graphics it produced.
As she progressed in school, the message from her professors and administrators changed. Students were encouraged to experiment with AI; For one assignment, Sato incorporated an AI-generated food truck into a logo design. She now regularly uses AI-powered tools like Photoshop’s “Erase” function to speed up her work, and she knows AI skills are in high demand.
He is still concerned about relying too heavily on AI but believes it cannot replace human precision or creativity. As a part-time designer at a clothing company this year, he offered to use AI to help with the mundane task of resizing product images. But he said the decisions he had to make in editing were too specific for the AI to understand. “You want something that’s based on that humanity,” Sato said. Sato has since accepted a full-time offer at the company where she interned.
The inevitability of AI has inspired new graduates like Tommy Lee to learn as much as possible about it. Lee, who graduated from Villanova University this month with a business degree, had not delved deeply into AI before taking a course on emerging technologies last fall. Since then, he estimates he has spent more than 800 hours experimenting on personal AI projects, learning from YouTube videos and trying out different models.
She automated her own job-application process, creating nine sub-agents to search for vacancies, prepare resumes, and fill out forms so she could focus on networking. (The system saved hundreds of screenshots of applications that it reviewed before submitting).
It paid off: Next month, he’s joining a soon-to-launch private-equity firm as an AI and systems analyst. Lee recently led a vibe-coding workshop (creating custom software with AI) in Philadelphia and hopes to conduct similar-style workshops for his new firm’s portfolio clients.
Some new graduate jobs will have more responsibilities than entry-level hires a few years ago. For example, Salesforce says it is recruiting and fast-tracking 1,000 AI-native graduates and interns this year for “practical, high-impact positions” in engineering, product, sales and other areas.
Day by day, they will delegate more administrative tasks to AI, just as 26-year-old Elizabeth Awad is currently doing in her role as a senior product manager. Awad, who recently completed a two-year entry-level rotational program, uses Slackbot, an AI agent, to organize her day and automate tasks. More time can be spent on strategic work by delegating meeting preparation and drafting product-requirement documents to AI, he said.
She recently led a demo on a Slackbot skill that formats messages in her writing style (it doesn’t capitalize words, unless she’s chatting with a senior executive). Within a week, she says, other project managers, designers, and engineers had cloned her repository to create their own agents.
The new expanded roles also mean new ways of training and overseeing graduate appointments. At SharkNinja, the company’s AI-native employees are working on high-level projects, like using the technology to synthesize data signals to make real-time adjustments to its supply chain.
Bosses are implementing morning and evening check-ins to help inexperienced employees course-correct and ask questions, CEO Barrocas said.
“We’re going to have to put some guardrails around them,” Barrocas said. “So, if you’re off the track, you’re off the track for a day, you’re off the track for a few hours.”
According to Strada’s survey of employers, working with AI tools has made critical thinking even more important than AI literacy.
Tim Walsh, the accounting firm’s US chairman and CEO, says AI is fundamentally reshaping work at KPMG, placing even greater emphasis on employee decision-making. This summer KPMG is piloting a new training program that focuses more on building critical thinking skills for its audit interns, including gamified exercises that force them to explore accounting scenarios by asking questions, avoiding bias, and using professional skepticism.
At the University of Vermont, business management professor Rocky DeWitt said she has moved from thinking about controlling her students’ AI use to using it as a tool to help them prepare for their careers. This spring, she asked Kanjorski and others to submit their chatlog history with each assignment, so she could evaluate how they interacted with the AI.
DeWitt then marked the conversation with feedback on quick-phrases, questions about what information he chose to leave out rather than what to include, and criticism about how he fact-checked the AI responses.
DeWitt said, “I wanted them to be able to explain in an employment interview how they are using technology as a discovery tool, and how they are creating value for a company.”
Kanjorski recalls a speech DeWitt gave to the class: “What you guys can do is go into these smaller companies… and bring similar solutions to them. You can be the person to lead that process.”
Write to Allison Pohle at allison.pohle@wsj.com and Roshan Fernandez at roshan.fernandez@wsj.com.