If you're a millennial like me who studied for their CPA years ago, then you might remember all the videos and audios and recorded lectures we ran through the weekend as we crammed by ourselves for hours and hours (usually with slow internet) before seeing the tired junior professor who'd stand in front of the classroom to help with any "doubts" students still had. While this approach definitely had its drawbacks, nonetheless it was a great way to scale up the classroom model where the best teachers cannot be present across classrooms, cities and even countries.
Fast-forward to the Gen Z era, where studying from online lectures, recorded videos, flipped classrooms, in-person classrooms and weekend study groups are the norm. Now, having said that, why do CPA pass rates remain such a
AI-CPA?
In this fast-paced era, where AI has taken the finance and accounting world by storm, watching just recorded lectures on what is ASC 606 and attempting multiple choice questions is like entering a Formula 1 race on a horse. It does the basic job, but it cannot keep up with the required pace. My last three years of doctoral study on opportunities and challenges for educators using AI in higher education have made me realize one thing: The struggle for teachers is as real as it is for students. But with a little creativity and AI, learning and teaching can be transformational for both teachers and students.
I recently took a global
This made me think of the ideal AI accounting learning management system and how it could make learning so much more interesting and adaptive through interactive videos that pause every few minutes to ask questions testing the student's level of concentration and understanding. Such a tool would be able to instantly generate easy, medium and hard questions depending on the competency level of the student. If a student isn't able to answer the question correctly, the adaptive AI model could reroute the student to easier-to-understand videos and then back to the main track. Now that's a very intuitive forward thinking learning experience.
No two learners are alike
Let's take an example of three scenarios: Student X understands leases, Student Y understands leases but struggles with journal entries, Student Z understands both but has trouble with simulations. All three students need different learning paths, which an AI LMS can identify and carve out before I can make a cup of tea.
For example, instead of a teacher explaining what an accounting memo is for an hour as a one size fits all scenario, the instructor could let AI do the "teaching" part and spend more quality time with students with an assignment that includes identifying weaknesses of an AI-generated accounting memo, verifying the citations, correcting accounting treatments or defending their conclusions, giving teachers more time and space to actively interact with students and assess their understanding levels. If students succeed, they learn. If they fail, they still learn.
Such an AI learning tool could potentially propel students forward by allowing them to study in "AI-personalized classrooms," even from remote villages in India or Africa, without ever losing the feeling of interacting with a teacher. It could also make it possible for students to study concepts in their native language, making foundational learning much stronger for non-English-speaking students. This could even help address the accountant talent shortage- but that is a topic for another day.
Should the teachers just retire then?
While some educators may raise concerns about this model, instead of looking at AI as a threat that would replace them, a CPA instructor now possibly has a very efficient teaching assistant, which can make teaching presentations (Gamma, Beautiful.ai), learning scripts (Claude or ChatGPT), interactive videos (Mindstamp or H5P), AI avatar videos (HeyGen or Synthesia), flashcards (Brainscape), mind maps (Napkin AI) or podcasts (NotebookLM) in mere seconds! The list is only growing.
With the amount of time saved using AI, teachers can spend more time with students identifying strengths and weaknesses using AI scoring dashboards and feedback tools. One of the biggest challenges for teachers today is the inability to identify students on the verge of dropping out of the CPA course. What if AI could read the signs of poor activity or understanding, and send warning signals to the teacher to intervene and help a student, instead of waiting until it's too late? When it comes to teachers, an important conclusion that surfaced from my doctoral research is that a single instructor cannot personally diagnose the needs of 500 students, but AI as a strong teaching assistant can potentially identify where those 500 students are struggling, allowing the human instructor to intervene where judgment and mentorship matter.
Right now, we're seeing AI being banned in a lot of educational institutions, based on claims that it kills a student's critical thinking and judgment. But what if we flip the scenario and actually use it to our advantage? After all, with AI taking up a lot of entry-level jobs, there is a risk of the learning curve getting eaten up by AI too, so it becomes necessary for students to develop higher-order skills.
Instead of fearing AI, students could harness its power and learn to recognize hallucinations, incomplete or inaccurate standards, citations or data bias. Despite the worries, AI can be a great solution if it is used deliberately with guardrails and academic integrity.
