In the experiment, a Harvard student, herself an AI enthusiast, enlisted eight professors and teaching assistants to evaluate essays generated by ChatGPT in response to authentic Harvard prompts.
The AI’s performance yielded impressive results, securing predominantly A’s and B’s, with only one C across social science and humanities subjects throughout the freshman year.
The overall grade point average (GPA) tallied to a respectable 3.34, according to the reports.The academics lauded the chatbot’s writing skills, commending its ability to craft sophisticated essays.
However, one reviewer advised the “student” to simplify their writing. Notably, the professors expressed concerns about the content and coherence of the AI-generated essays, criticizing the arguments for being “consistently large and unclear.”
The increased popularity of generative AI, sparked by the release of ChatGPT, has given rise to accusations of academic dishonesty, though not all have been proven true.
- Advertisement -
Similar experiments, albeit on smaller scales, have been attempted by professors at various institutions, highlighting the challenge faced by colleges in dealing with this technological advancement.
In one such earlier experiment, Wharton professor Christian Terwiesch tested an earlier version of ChatGPT powered by GPT-3.5 using questions from a final exam. The chatbot’s performance on the exam was relatively modest, earning a B- grade.
Consequently, colleges have grappled with the consequences of this new technology, causing some professors to take proactive measures.
The emergence of an atmosphere of distrust between students and lecturers has further complicated matters.
In response, the higher education sector is working on developing guidance and policies to address the surge of generative AI.
Initially met with resistance and outright bans, some colleges are now exploring ways to integrate and harness generative AI tools. Efforts are being made to capitalize on the potential benefits while safeguarding academic integrity.
As institutions strive to strike a balance, the responsible implementation of AI tools may pave the way for a harmonious coexistence between technology and academia.