AI still isn’t making a serious impact on university education – here’s why

Artificial intelligence (AI) promised to revolutionise the field of education. But how, and how much, is it actually being used by students and teachers today? Our recent article sought to answer this question by analysing AI’s integration into universities from two perspectives: scientific (how it is actually being used) and social (perceptions of its use).

While researchers tend to highlight the opportunities and challenges AI presents for personalised learning, social perception tells a different story, reflecting a slower and more uneven adoption, particularly in Europe.

Personalisation, virtual tutors and admin

In recent years, AI has been steadily rolled out in education, particularly in areas such as personalised learning, virtual tutors and the automation of administrative tasks.

Platforms such as Smart Sparrow, Knewton, Century Tech, and Khan Academy use AI to adapt the pace and content of learning to students’ individual needs. These systems use AI to analyse student performance (correct or incorrect answers, response time, error patterns), as well as to automatically adjust the level of difficulty, type of content or pace, and to recommend additional exercises, videos or reading.

Tutoring systems are another use. These are chatbots or virtual assistants that interact with students as a human tutor would. Their typical functions are to answer questions about content, propose exercises, give explanations, solve problems step by step, motivate and accompany the student in their progress.

Some examples are Khanmigo (Khan Academy + GPT-4), which helps with maths, writing and science, Duolingo Max, an AI for personalised language tutoring, Google’s Socratic, which answers questions with visual explanations, and Carnegie Learning’s Mika, an AI-powered maths tutor.

All of these platforms use machine learning models that recognise students’ strengths and weaknesses. This technology is already being used successfully in disciplines such as medicine, electronics and linguistics, where large-scale data analysis and automation are essential for teaching and research.

This technology also allows for the efficient management of administrative processes, such as exam correction and tracking grades and performance.

Read more:
AI in universities: How large language models are transforming research

Limitations and challenges

Despite the expectations surrounding AI, its impact on university education remains modest. Globally, its use in universities is still in its infancy, and AI adoption varies greatly between regions and disciplines. While significant progress has been made in some areas, such as health sciences, other areas like the humanities are only just beginning to explore the possibilities of this technology.

One of the main challenges is a lack of training for teachers and administrators in the use of AI tools. Many teachers do not have the necessary skills to integrate them into their classrooms, which limits their adoption. In addition, the lack of clear policies on student data privacy and the ethical use of this technology poses significant barriers.

Read more:
AI is now part of our world. Uni graduates should know how to use it responsibly

Europe lags behind

Despite being a leader in regulations on the ethical use of technology, Europe is lagging behind in terms of scientific research on how to integrate it into educational methods.

There are exceptions, including the UK, which has very solid research on the ethics of educational AI and adaptive teaching models and automatic assessment. In Germany and the Netherlands, European Union interdisciplinary projects are also combining education, cognitive sciences and computer science.

The United States leads in scientific publications, patents, and the development of educational technologies based on AI. Meanwhile in China, there has been a significant increase in publications and applications of educational AI, especially in adaptive learning and facial recognition in smart classrooms. Massive state investment in “Smart Education” is part of its strategy for leadership in AI.

Latin America (especially Brazil, Chile, and Mexico) also has a growing research output, with advances in adaptive educational platforms and learning data analysis. There is a growing interest in using this technology to reduce educational gaps, and in improving access to it in disadvantaged areas.

Read more:
AI chatbots can boost public health in Africa – why language inclusion matters

Social perception: indifference at best

In our recent study, we compared research on the use of AI in education with the social perception of its use through mentions on social media.

Use of AI in education has sparked social debate, but we found that social media users generally are generally neutral, or even unaware, when it comes to its impact on universities. Despite its growing presence in conversations about educational technology, most online mentions of AI in education express neither enthusiasm nor significant concern.

While researchers focus more on the development and academic impact of AI, social media users mainly focus on AI tools like ChatGPT, which help students with more practical, everyday tasks.

Where is AI in education headed?

In terms of what this technology can offer, personalised learning and task automation are just the tip of the iceberg. To harness its true potential, it is crucial to invest in teacher training, develop clear policies and promote greater collaboration between researchers, educational institutions and society.

AI is opening new doors in education, but its adoption still faces significant barriers, especially in Europe. Despite advances in medicine, electronics, and linguistics, its widespread implementation in other areas requires both scientists and society to bridge existing gaps, and make the most of its opportunities.

_{Images are for reference only.Images and contents gathered automatic from google or 3rd party sources.All rights on the images and contents are with their legal original owners.}