LLMs in education: from crutches to jumping pole
You see students using ChatGPT to finish tasks fast. Professors and teachers rant about the rate of pupils using large language models (LLMs) to turn in an essay. Maybe even you have asked an artificial intelligence (AI) for help in a task you were supposed to do yourself. LLMs in education are here to stay. Regardless of how we feel about it, they are available everywhere. But right now, that is not a good thing. Having an educational model set up without LLMs only benefits those who can use AIs like ChatGPT as a shortcut to do less work. And yet, we face the chance to use those very AIs to improve education instead. How to turn crutches into jumping poles We should not pretend that the task ahead is easy. When slides were invented, shifting from blackboards and books to presentations took time. In fact, these tools remain in the classroom, and that is well. The same happened with Google and the internet. Information at the student’s fingertips made it so schools and universities had to adapt. And while it is still a work in progress, these technologies have been adopted. The same is happening with AI and LLMs. For example, Estonia is already investing in plans to equip the classroom with these tools. And the European Union (EU) is forming teachers in the use of AI for education in STEM. And yet, technological advancements can increase disparity in the classroom. AI as an opportunity to bridge educational gaps When it comes to STEAM education (Science, Technology, Engineering, the Arts, and Mathematics), there are still gaps and barriers for many students. First, we have a socio-economic barrier. Students with less means are at a clear disadvantage in the current education system. Then, students in rural areas may have to move or commute for studies, resulting in economic difficulties and less support. They also may have fewer resources, especially technological ones. And there is still a gender gap in STEAM. Women are fewer than men in many areas, especially in decision-making roles. With the introduction of AI in the classroom, we can increase these differences if we are not careful. But AI can also let us promote a more equal, balanced, and empowering education system. This is exactly what STEAMbrace stands for and aims to do. Promoting a fairer education system STEAMbrace is pushing to create the STEAM Alliance for Europe. This is a European-wide collaboration. It involves educators, policymakers, cultural and creative industries, civil organizations and research organizations. The aim? To build on previous European initiatives and use creative approaches to make STEAM mainstream and accessible for everyone. The main focus is on the gender gap, reducing socio-economic barriers, and boosting rural education to reduce disparity. AI in the classroom to reduce disparities. LLMs are often trained on public data. If the data itself is skewed, the LLMs show prejudice. This is already happening, for example, they show a gender bias. We need to do a better job if we want to promote gender equality. This means providing students with unbiased LLMs they can trust. The use of AI and LLMs also requires an internet connection, and paying means you get access to a better version of, for example, ChatGPT. This poses an issue for rural students with slower internet access, if any. It also benefits members of higher socioeconomic layers. To reduce these disparities, we must provide students with the right tools and opportunities, both in the classroom and beyond. LLM-conscious teachers and students To achieve this, we must push for LLM-conscious teachers and students. But it takes a village. We need the effort of many other groups. Researchers, to identify possible gaps and the best educational practices. Policymakers and administrators, to give the teachers the needed training and tools. And educators, to provide students with the knowledge to use AI and LLMs responsibly. Critical thinking is also key to interpreting the results of AI use. Conclusion STEAMbrace pushes for the formation of an EU alliance that can tackle current educational issues. These include the gender gap, socio-economical differences and rural disparities. AI can contribute to deepening these issues or to remedy them if used properly. The STEAM Alliance for Europe can be used to make sure that LLMs and AI are implemented in the classroom with the right mindset. By informing the participants involved about the potential issues that may come from using LLMs, we can be ready to use them more safely and reduce disparities.