Human cognition and AI-generated texts: ethics in educational settings

Abstract

As artificial intelligence (AI) tools refine their ability to produce increasingly coherent and fluent texts, the ethical implications of their utilization in educational contexts warrant discussion. Upon briefly overviewing recently emerging evidence on AI-generated and human-written texts, I note that detecting AI-generated texts may be difficult not only for humans but also for technological detectors and classifiers. Accordingly, for the investigation of the relationship between human cognition and the recent developments in AI language generators, I suggest that the main focus should be on educating students on using AI only in ways that are conducive to higher-level thinking and learning, more so than how students' AI-generated texts can be better detected. To achieve this, educational curricula should adapt to today's AI revolution by starting to teach students how to utilize AI to support their analytical and critical thinking skills, rather than using AI tools to do their work for them and thereby letting such tools erode these qualities. For this to be possible, students should be motivated to use AI only to support their thinking and learning, which can be possible through the creation of an honor code specific to the ethical utilization of AI in educational contexts. As this AI-specific honor code does not yet exist, I outline ways based on previous honor code research for this code to be created and such a culture of AI ethics to be implemented. In sum, I recommend ethical and practical measures that can facilitate a healthy AI-human partnership in educational settings.