by Keith Quesenberry*

Artificial Intelligence (AI) is one of the most disruptive forces in both professional careers and higher education. As discussion of AI heats up and we debate whether future frontier models endanger society, current AI tools are being used today in businesses across the country. How does this technology impact students entering the workforce? Are entry-level jobs declining? How will those jobs change and what does that mean for business and education long-term?

The latest evidence from the Stanford Digital Economy Lab analyzing ADP payroll data finds no evidence of widespread economy-wide job displacement. In fact, entry-level hiring is up 5.6% based on data from the National Association of Colleges + Employers (NANCE). However, the nuance in the Standford data is that young worker employment in AI-exposed occupations is 19% below less-exposed occupations. This gap doesn’t appear with more experienced workers.

The declines are concentrated in occupations where tasks are susceptible to Large Language Model (LLM) automation versus LLM augmentation. This tells us that AI is not taking away all entry level jobs but that it is slowing hiring in jobs that focus on skills AI can easily replace – typically lower-level, route tasks.

A first concern is we need to ensure we shift what we’re teaching students to better prepare them for these evolving job skills. As lower-level tasks are replaced with AI automation, students will need to enter the workforce with higher-level strategic skills that enable them to guide and judge AI tools as augmentation. In practice this may feel like students entering their careers up a rung on the ladder. The latest NANCE report reveals the majority of employers assign interns projects that require AI tools and demand for AI skills in entry-level jobs nearly tripled since Fall 2025.

Yet, we can’t jump to all AI automation of these tasks in the classroom. Professors must navigate what this means in their individual disciplines. As we attempt to balance AI tool use that meets industry expectations, we must also acknowledge that types of AI use can hinder learning undermining the higher-level strategic skills students will need. Certain things simply require the hard work and friction of learning on your own to internalize the knowledge you may later use to guide and judge LLM automation.

A second concern is that no matter how much we prepare students in the classroom, certain aspects of a career are learned on the job. This is why we emphasize the value of internships. But even beyond an internship a lot of learning happens in the first-year on the job. Most of that learning happens in the small, more menial, route tasks. The tasks may be menial but that doesn’t mean they are meaningless. What happens if first-year employees are no longer given those smaller tasks because AI is doing more of them? If that happens students may have to enter their careers on a higher rung and that rung may be hollow because it no longer prepares them for the more advanced jobs the way entry-level jobs used to.

These concerns present immediate and longer-term challenges for industry and academia and the solution doesn’t rest on either. In education, we can strive to bring more of those on-the-job learning activities into the classroom. Experiential learning includes case studies, real client work, and simulations, but it also can be more frequent low-stakes opportunities to be wrong and receive feedback like a manager or mentor.

We can also lean into the uniquely human skills that have traditionally been labeled as soft skills and deemphasized for technical preparation. Both are important: especially in the age of AI. I constantly remind my management students of the importance of the skills they learn in their humanities classes as part of a strong liberal arts foundation.

In industry, organizations may need to invest in first-year knowledge workers the way many trades have apprentice programs. Ensure younger employees are taught by experienced workers on the nuances of manual tasks to be prepare them to guide and gauge automation. If not, I worry organizations may end up with workers who appear more capable than they really are because their most visible work is being done by AI automation. Then when they enter higher-level positions, they are not prepared to take the reins. Recently we’ve seen companies that have rushed into AI automation taking a step back realizing that it is more complicated than simply replacing tasks.

The immediate impact on entry-level jobs doesn’t show widespread impact, but it does show a gap in certain occupations that could widen. The traditional pathways to learning a profession may be changing. Now is the time for education institutions and business organizations to plan ahead and meet the evolving landscape.

**Keith Quesenberry is associate professor of practice in markets, innovation & design in the Freeman College of Management at Bucknell University.