AI Skills Now Matter More Than Degrees: Employers Put the MBA on Notice

For many U.S. finance employers, AI skills now outweigh an MBA. A new PwC survey of more than 1,000 financial-sector executives finds 86 percent consider AI training for new hires more valuable than an MBA, and 91 percent boost pay for staff with AI expertise.

That shift doesn’t just pressure traditional management programs. Universities must also decide what students should master without AI—and when generative AI should be built into the curriculum. A striking case at Brown University shows why that debate is becoming urgent. As reported by Forbes.

PwC sees a clear pivot in finance from credentials to practical AI know-how. Among respondents, 91 percent say employees with AI skills can earn more. In addition, 62 percent want to hire new staff with specific AI expertise, and 61 percent plan to upskill or reskill current employees.

This goes far beyond using a chatbot. Financial institutions want people who understand how to deploy, oversee, and integrate AI systems and AI agents into business processes. AI agents are software systems that can independently execute multiple steps to achieve a goal.

That doesn’t mean degrees are suddenly worthless. It does show employers are increasingly judging candidates by what they can do with technology—not just the diploma they hold.

The distinction matters. An MBA still delivers strategy, finance, leadership, and organizational insight, but AI skills evolve faster than traditional programs can typically adapt. That opens space for shorter programs and certificates tightly aligned with new technology.

Major institutions are responding. In April 2026, Wharton launched the online certificate Digital Strategy in the Era of AI, training professionals to deploy AI and digital tech strategically across organizations.

Wharton also offers a Digital Leadership Certificate, including the course Artificial Intelligence for Business. The full program currently costs $3,000 and can be completed fully online.

That creates an alternative model alongside traditional degrees. Professionals can buy targeted AI expertise without committing years to a full program.

For universities, it’s both opportunity and threat. They have faculty, brand, and deep expertise—but increasingly compete with compact learning paths tailored to the skills employers want right now.

While employers reward AI fluency, universities struggle with students using the same tools during assessments.

The issue surfaced in an economics course by Professor Roberto Serrano at Brown University. In a take-home midterm, the average score hit 96 percent. Forty students reportedly earned a perfect score. Serrano suspected many had used generative AI.

He and his graders then fed the questions into ChatGPT. According to Serrano, several chatbot answers closely matched student work—in methods and in wording. Inside Higher Ed notes, for example, a proof where a direct argument was obvious, yet both ChatGPT and multiple students used a more cumbersome proof by contradiction.

Serrano responded by holding an in-person final. If results were similar to the midterm, the original grades would stand.

They weren’t.

The in-person exam average fell to 48.6 percent—the lowest Serrano says he’s ever seen for the course. Eighteen students dropped the class before the exam, nine registered students skipped it, and nineteen ultimately failed the course.

Those figures don’t prove at an individual level who used AI. But they highlight how hard it’s becoming for institutions to verify what students truly know when powerful generative AI is available outside the classroom.

Brown is now treating the issue as broader than potential exam fraud. In July 2026, the university released a report on generative AI in teaching and learning, based on feedback from nearly 700 students, faculty, and staff, and an analysis of almost 3,000 syllabi.

A striking finding: over half of the syllabi reviewed had no explicit policy on generative AI. Where rules did exist, they ranged from full permission to outright bans.

Brown is now drafting university-wide ground rules. A new committee will create templates for courses where AI is fully banned, partially allowed, or deeply integrated.

That approach underscores a reality: simple bans are becoming untenable. Graduates enter workplaces that expect AI fluency. At the same time, universities must prove that students can still demonstrate core skills without leaning on a language model.

The likely shift isn’t the end of diplomas—but a change in what they certify.

Traditionally, a diploma signals two things: completion of a program and evidence of specific knowledge and skills. Generative AI complicates that second claim, since students can produce work—assisted by AI—that surpasses their independent level.

Yet walling off AI from education makes little sense when industry uses it daily. Brown acknowledges the tension: students already use generative AI widely, while both students and faculty worry about its impact on critical thinking, learning, and academic integrity.

Expect institutions to measure two skill sets separately more often: independent understanding and performance with AI.

For example, a student might prove foundational knowledge in a proctored written or oral exam. In other assignments, the same student could be graded on prompt design, validating AI outputs, spotting errors, and responsibly integrating results.

That blend is closer to the job market. Employers don’t need people who can reproduce everything without AI but can’t use modern tools. Nor do they want employees who rely entirely on a model and fail to notice when it’s wrong.

The PwC survey and Brown’s moves point to the same structural shift from different angles. Employers are elevating verifiable AI skills, while universities must redefine the human expertise a diploma should guarantee.

For students, that reframes the ROI question. A long, expensive degree still pays off if it delivers deep domain knowledge, networks, and proven expertise. But if a program mainly certifies knowledge that AI can easily reproduce, pressure mounts to redesign it.

The strongest diplomas of the AI era won’t try to prove graduates can work without technology. They’ll show that students understand their craft, can reason independently, and can wield AI more effectively than someone without that subject-matter depth.