The groundbreaking AI tool, ChatGPT, has been generating interest since 2022, with its use in academic settings now under study. Researchers recently compared the essays written secondary school students with those generated ChatGPT, finding that the AI outperformed students in all criteria, particularly excelling in language mastery according to an article TechXplore.
ChatGPT faced an initial setback with version 3.5, but its successor, version 4, managed to make significant progress, nearly six months later, in a German secondary school test. The study titled “A large-scale comparison of human-written versus ChatGPT-generated essays” researchers at the University of Passau, investigated the impact of AI-generated content on the educational system.
The study revealed that both versions of ChatGPT outperformed students, with GPT-4 achieving the best score. This outcome points to the potential benefits of employing AI tools in schools. The study was an interdisciplinary effort involving computer scientists and experts from the field of computer linguistics and computer science education.
Teachers were also involved in evaluating the essays based on the Ministry of Education of Lower Saxony guidelines. GPT-4 scored 5.25, GPT-3 scored 5.03, while students averaged 3.9 in language mastery according to 111 participating teachers.
The study, published in Scientific Reports, raises thought-provoking questions about the potential impact of AI-generated language on human communication. While it does not suggest poor English language skills in students, it does highlight the exceptional performance of AI models and the need to consider how exposure to machine-generated language may influence and shape human language in the future.