Why do we need to employ exemplars in moral education? Insights from recent advances in research on artificial intelligence

ABSTRACT

In this paper, I examine why moral exemplars are useful and even necessary in moral education despite several critiques. To support my point, I review recent AI research demonstrating that exemplar-based learning is superior to rule-based learning in model performance in training neural networks, such as large language models. I particularly focus on why education aiming at promoting the development of multifaceted moral functioning can be done effectively by using exemplars, which is like exemplar-based learning in AI model training. Furthermore, I discuss the potential limitations and issues related to exemplar-applied moral education with findings from recent AI research raising concerns about model biases and toxic outcomes. I propose ways to address the concerns regarding employing moral exemplars as well. As remedies, I suggest that autonomy-supporting deliberative and reflective learning should be utilized. Furthermore, based on the discussion, I examine how macroscopic socio-cultural aspects influence the effectiveness of exemplar-applied education.

KEYWORDS:

• Moral education
• moral exemplar
• artificial intelligence
• neural network
• large language models

Acknowledgments

I thank the handling editor, four anonymous reviewers, and Sora Kim for their invaluable feedback on the earlier version of the manuscript.

DISCLOSURE STATEMENT

No potential conflict of interest was reported by the author.

FINANCIAL STATEMENT

I confirm that I have no relevant financial interests in relation to this publication.

Notes

¹ Emulation within the context of moral education is different from mere imitation (Fridland & Moore, 2015; Kristjánsson, 2006). Mere imitation is about immediately copying a presented behavior, so it does not ensure the implementation of such a behavior across different situations and contexts. On the other hand, emulation is more about learning mechanisms and rules from presented models, develop one’s own generalized behavioral strategies, and finally, behave in such a way even in different environments and situations.

² The blueprint of flourishing is about the extent to which one understands the importance of flourishing and how such understanding motivates and directs their action. Moral sensitivity is related to one’s ability to detect morally salient aspects of a situation and figure out the best solution. Reason-infused emotion deals with regulating one’s emotions for optimal emotional experiences with guidance of reasoning., Moral adjudication is about balancing different virtues and strengths to be able to render a best decision in a situation while addressing conflicts.

³ I plan to discuss further details about the implications of meta-learning in AI research in moral psychology and moral education. However, that is out of the scope of the current paper, so that will be addressed in another paper.

⁴ E.g., “Corgi has six squeeky puppets. She buys three more boxes of puppets. Each box has two puppets. How many squeeky puppets does she have now?” “Corgi started with six squeeky puppets. Three boxes of two puppets each is six puppets. 6 + 6 = 12. The answer is 12.”

⁵ E.g., “Corgi has six squeeky puppets. She buys three more boxes of puppets. Each box has two puppets. How many squeeky puppets does she have now?” “The answer is 12.”