Large Language Models and the value of education

Large Language Models (LLMs), such as ChatGPT or LaMDA are often presented as a breakthrough in AI because of their ability to write convincing texts in a matter of seconds and when presented only with a short opening sentence. In education specifically, there are concerns about LLMs displacing the opportunities for critical reasoning through the act of writing and the consequent pedagogical challenges in this regard. In this presentation, I explore the critical appropriation of LLMs in higher education. Theoretically, I will follow the lenses of technological mediation and sociotechnical systems approach to explore the potential impact of LLMs on the value of education. First, I will survey the technological background of LLMs and the students' emerging reactions to this technology before proceeding with the ethical analysis. Here, I will focus on the appearance of intelligence, misinformation risks, bias, and the ecological sustainability aspects of using LLMs in the classroom. The paper will draw on a case study of using ChatGPT in philosophy-related classes and other empirical data on using LLMs in education. I will not only focus on concerns but also present a range of opportunities for the educational context, as well as suggest several ways of how to approach LLMs in higher education from a perspective of human-AI collaboration and sustainable integration.