Evaluate my AI teacher? Students’ perceptions of chatbots will influence how they learn with AI

A “transformation” is upon us. After a succession of several years of educational technology products that once promised to shake things up, it is now AI’s turn.

Global organizations like the Organization for Economic Co-operation and Development, as well as government agencies, are promoting AI to the public as “transformative.”

Leading AI companies with large language model (LLM) chatbots offer “education-focused” products, such as ChatGPT Education, Claude for Education, and Gemini in Google for Education.

AI products facilitate exciting new ways to search, present and interact with knowledge and have sparked widespread interest and enthusiasm for technology among young learners. However, there are critical concerns surrounding the use of AI, such as data privacy, transparency, and accuracy.

Current conversations about AI in education focus on the idea that it will disrupt teaching and learning systems in schools, lesson planning and grading by teachers, or individualized learning (for example, through personalized tutoring of students with chatbots). However, it remains an open question when and if AI will transform education.

In the meantime, it is essential to think about how student engagement with chatbots should lead us to examine some fundamental assumptions about human learning.

Learning is a social affair

How students perceive their teachers and their own ability to think (called metacognition) are extremely important for learning. These factors need to be considered when we think about learning with chatbots.

The popularity of the Rate My Professors website in Canada, the United States and the United Kingdom speaks to the importance of what students think of teachers.

With the incursion of AI into education, students’ conception of their AI tutors, teachers, and evaluators will also be important for multiple reasons.

First, learning is a profoundly social affair. Whether it’s how a child learns by imitating and modeling others, or by engaging or being influenced by peers in the classroom, social interactions are important in how we learn.

As chatbot usage reaches over 300 million monthly users, conversational interactions with LLMs also represent a new space for parasocial interaction for people around the world.

What we think about interaction partners

Second, theory of mind frameworks suggest that what we think about others influences how we interact with them. How children interpret, process, or respond to social cues influences their learning.

To develop this idea further, beyond other students or teachers as interaction partners, what we think about learning tools has an influence on how we learn.

Our perception of tools and their means – the quality or property of a tool that “defines its possible uses or clearly indicates how it can or should be used” – can have consequences for how we use the tool.

Perceived means can dictate how we use tools, from utensils to computers. If a learner perceives a chatbot as being adept at generating ideas, then this could influence how they use it (for example, for brainstorming or editing).

New “social entity”

At a minimum, AI systems represent the entry of a new social entity into educational environments, as they have done in the social environment. Popular conceptions of AI can be understood within the broader framework of a theory of artificial minds, referring to how humans infer internal states of AI to predict actions and understand behavior. This theory extends the notion of theory of mind to non-human AI systems.

A person’s theory of artificial minds could develop based on biological maturation and exposure to technology, and could vary significantly between individuals.

Three aspects to consider

It is important to consider the impact that students’ conceptions of AI may have on trust in information received from AI systems; personalized learning using AI; and the role that AI can have in a child’s social life:

1. Trust: In human learning, the judgments we make about knowledge and learning contribute greatly to the acceptance of the ideas inherent in the learning material.

Based on recent studies of children’s interactions with conversational AI systems, we find that children’s trust in information from AI varies depending on factors such as age and type of information. A learner’s theory of artificial minds would likely affect their willingness to trust information received from the AI.

2. Personalized learning: Research on intelligent tutoring systems (ITS) has shown excellent results on how traditional ITS, without the engagement of a chatbot, can provide structure to learners while helping them identify gaps in learning for self-correction. New chatbot-based ITS, such as Khan Academy’s KhanMigo, are marketed as providing personalized advice and new ways to interact with content.

A learner’s theory of artificial minds could affect the quality of interactions between them and their AI chatbot tutor and the extent to which they accept their learning support.

3. Social relationships: The artificial friend (“AF”) in Kazuo Ishiguro’s “Klara and the Sun” is a poignant literary example of the impact an artificial entity can have on a growing child’s sense of self and relationship to the world.

We can already see the detrimental effects of introducing AI social chatbots to children with the tragic suicide of a child who was allegedly engaged in emotional and sexual conversations with a Character.AI chatbot.

Social relationships with AI involve a serious renegotiation of the social contract regarding our expectations and mutual understanding. Here, relationships with children require special attention, above all if we want children to develop social relationships with AI.

Where do we go from here?

Many discussions are currently taking place on the mastery of AI, involving for example the understanding of how AI works, its limits and its ethical issues. Throughout these conversations, it is essential that teachers recognize that students have an intuitive sense of how AI (or a theory of artificial minds) works. Students’ intuitive sense of AI shapes how they perceive its educational possibilities, even without formal learning.

Teaching must take into account students’ cognitive development, existing experiences, and changing social contexts.

The “rate my AI teacher” future is coming. A focus on students’ conceptions of AI will be necessary to ensure effective, ethical and meaningful integration of AI in future educational environments.

This article is republished from The Conversation under a Creative Commons license. Read the original article.