Artificial Intelligence tools like Copilot and ChatGPT can increase productivity. They speed up workflows and can be used to automate freeing people to spend their time on creative tasks. Advances in artificial intelligence technologies had a significant impact on the economy in 2023. Fortune magazine attributes AI-related economic growth as having helped “saved the economy” because growth in this sector offset disappointing returns in other sectors. 

But with great power comes great responsibility. The tools of AI allow freshman college students to do what a year ago would have required a decade of professional experience. Educational institutions must carefully consider where and how these tools can be used without undermining the quality of education. Otherwise, overreliance on artificial intelligence will make us, and the next generation of professionals, artificially smart but lacking in basic understanding.

Navigating Challenges

Universities are struggling to find solutions to widespread cheating facilitated by ChatGPT, Copilot, and other tools. This phenomenon is occurring worldwide. One mitigation technique that may surprise technology professionals is the prevalence of low-tech solutions. Many universities are returning to written in-class exams. Curtin University of Dubai instituted a return to written exams after a third of students in one of its classes was caught using ChatGPT to write their class essays. Business Insider found many professors considering returning to oral exams as a surefire way to test for understanding in students. These solutions ensure that students understand the content covered in their classes, but they are time-consuming and require increased effort from professors. Additionally, it is much more difficult to outsource grading to a teaching assistant, because the level of conceptual mastery required to accurately proctor an oral exam is greater than that required to compare an exam to a solution key. Unless technology is developed to accurately detect plagiarism, these costs may be prohibitive for universities with high faculty-to-student ratios. Despite the challenges, serious effort needs to be made to ensure that understanding remains the metric by which students are evaluated.

Sadly, universities have been moving away from this standard for some time. As of 2023, over 80% of colleges and universities in the United States do not require ACT or SAT scores for college admission. Often, they won’t even be considered. Perhaps there are flaws with standardized testing, but moving to an application based solely on GPA and an admission essay is a mistake. Oral examinations are rarely applied in engineering undergraduate programs. Engineers have a reputation for on average being awkward. If presentations and oral examinations were metrics determining the success of students, then much of this could be alleviated. Tragically, these trends are also prevalent in graduate and PhD programs. I’ve observed several programs that are moving away from requiring oral examination components for qualifying exams. If nothing else, surely the entrance examination to a PhD program should contain hard metrics requiring students to demonstrate existing mastery of content without technological aid.

Solving Problems Requires Creativity and Understanding

If steps aren’t taken to ensure that conceptual mastery remains the standard for higher education, most students will complete their degrees as technicians. They will be able to apply existing technology to complete a low-level technical job by relying on tools and processes built by engineers. Universities should produce graduates able to create new solutions that solve complex problems, and thus require deep understanding, creativity, and problem-solving ability. Without such understanding, these graduates are ripe for automation. 

The advent of ChatGPT and other generative artificial intelligence tools serves as an opportunity to reassess education methodology. By focusing on conceptual mastery, higher education can produce graduates who can use both their deep understanding and these exciting tools to solve cutting-edge technical problems and create real-world solutions. If not, there will come a day when our professionals are artificially smart.