For the past few years, the keynote stage at many conferences has been dominated by a single, looming specter: artificial intelligence (AI). We’ve all likely seen the demos of large language models drafting technical memos in seconds, automated bots reconciling thousands of transactions almost instantly, and heard predictions that the entry-level accounting and finance professional may soon become obsolete.
As certified public accountants (CPAs), we’re trained to look at the data, assess evidence, and evaluate risk. If we look only at the efficiency metrics, we might be tempted to believe that the internship—the traditional “boots on the ground” role for junior-level staff—is quickly vanishing. Why pay a college student to spend hours “ticking and tying” when a software plugin can do it in milliseconds? This perspective isn’t just short-sighted—it’s a fundamental misunderstanding of what our profession is and how a CPA is made.
AI may reduce the need for interns to perform repetitive manual work, but it also increases the need for early-career professionals who can validate outputs, understand source documentation, ask better questions, and develop the professional judgment required of a CPA throughout their career. That’s why the human intern remains a critical part of both the profession’s talent pipeline and its risk-management framework.
One of the core tenets of our profession is an attitude that includes a questioning mind and a critical assessment of evidence (i.e., professional skepticism).
AI isn’t skeptical—it’s predictive. It identifies patterns, generates responses, and produces outputs that may appear polished and authoritative. But AI doesn’t know whether its answer is true; it predicts what answer is likely to sound right. It doesn’t understand the concept of fraud, nor can it sense when a client’s explanation doesn’t match the numbers. An intern, however, has the capacity for intuition. When an intern is tasked with testing a sample, they aren’t just checking a box. They’re learning what “normal” looks like. They’re developing the judgment, curiosity, and discipline that eventually become professional skepticism.
If anyone doubts the danger of unsupervised AI, we can simply look at a recent example from within the legal profession for a sobering reality check: The prestigious, global Wall Street law firm Sullivan & Cromwell recently issued a public apology for submitting a document to a federal court that contained citations generated by AI. These citations weren’t just wrong—they were “hallucinations” of entirely fabricated cases and code sections that appeared authentic but didn’t actually exist.
This incident serves as a warning for the accounting profession. If a law firm with immense resources can fall victim to AI fabrications because they lacked the human oversight to verify the work, what does that mean for a tax return or an audit engagement?
In our profession, that human oversight has traditionally been the intern or the staff associate. They’re the ones who cross-reference work papers, verify source documents, and ensure that the summary matches the reality of the ledger. The Sullivan & Cromwell case proves that you can’t simply replace a junior professional with a software license. You need human eyes trained in the fundamentals to act as the final guardrail against digital fiction.
There’s also a practical reason why firms and finance departments need interns now more than ever. Many partners, managers, controllers, and chief financial officers are still learning how AI tools fit into their workflows. By contrast, today’s accounting students are already using these tools in the classroom and in their daily lives. Of course, that doesn’t mean interns understand the full complexity of an audit, tax engagement, financial close, regulatory filing, or internal control environment—they don’t. But they do bring fluency with emerging tools, comfort with experimentation, and a willingness to ask why a process has to be done the way it’s always been done. In that sense, interns can become digital accelerators. They bring fresh perspective to organizations that may be struggling to modernize. They can help test workflows, identify inefficiencies, support data cleanup, and contribute to responsible experimentation with technology.
Yet, at the same time, firms must be intentional with how they utilize interns and their capabilities. The solution isn’t to give interns unrestricted access to AI tools or sensitive client data. The solution is to redesign internships so students can learn how to use the technology and apply professional judgment around it.
Fortunately, moving the accounting internship away from repetitive task completion to structured professional development can be achieved by designing internships around these four pillars:
Additionally, as firms redesign internship programs, they must be clear about the boundaries. Interns should learn not only how to use AI but when not to use it. Client confidentiality, proprietary data, independence, regulatory expectations, cybersecurity, and ethical obligations must be built into any AI-enabled internship experience. An intern should understand that just because a tool is available doesn’t mean it’s appropriate for every task. They should be trained on firm-approved platforms, data-use restrictions, review protocols, and documentation standards.
This is also an opportunity. If we teach students early that technology must be paired with ethics, governance, and accountability, we’ll prepare them to become the kind of professionals the future requires.
The argument for replacing interns with AI rests on the flawed premise that an intern’s primary value is manual labor. Historically, interns have performed many repetitive tasks. They footed schedules, scanned invoices, tied numbers, rolled forward workpapers, and completed checklists. But the true value of an internship has never only been the work produced. The deeper value is the evolution of the practitioner. An internship is a clinical rotation for a future CPA. It’s where students begin to see how classroom concepts become professional responsibilities, where evidence supports conclusions, deadlines affect judgment, teams operate, clients communicate, and ethical obligations show up in real time.
AI can process information, but it can’t become a CPA. If we want future CPAs who can exercise judgment, challenge outputs, protect the public interest, and lead technology-enabled finance functions, we can’t automate away their first training ground. Instead, we must redesign it.
The internship of the future should be more analytical, technology-enabled, ethical, and engaging than the internship of the past. It should expose students to AI but also teach them to question it. It should reduce mindless work but preserve the foundational experiences that teach accuracy, accountability, and skepticism.
Although AI can give us quick answers, only a trained human professional can determine whether those answers are complete, appropriate, and reliable.
