Description

There are a number of ethical issues to consider regarding the educational use of Artificial Intelligence (AI/AIED). These include the collection, management, interpretation, ownership, security, and use of data and information generated through the use AIED technology, the design, development, use, and evaluation of generative AI and AIED technology in teaching, the informed and equitable participation in the use of AI and AIED technology in teaching and learning, and the threat of the displacement of professional and academic autonomy, authority, and agency and their protection.

With these issues in mind the purpose of this presentation is to engage participants in an interactive, deliberative process of reflection and dialogue concerning possible principles for the ethical use of AI in Education as responses to these issues.

This reflection and dialogue will utilize the normative methodology of “Reflective Equilibrium”, which is a prominent method of ethical inquiry (based upon the work of John Rawls, Thomas Scanlon, and Cass Sunstein, among others). Reflective equilibrium seeks the consistency/coherence between our considered judgments and ethical principles upon due reflection. We consider whether our principles are consistent with our considered judgments upon due reflection. Considered judgments constitute ethical fixed points, in the sense that they are convictions that we are least likely to give up. Our considered judgments are, however, dynamic, not static, in the sense that their consistency with principles is subject to revision and vice versa. In that sense they are “provisional”.

The presentation will ask participants to identify their considered judgments and then to reflect upon and discuss in groups the consistency between those judgments and possible principles for the ethical use of AI in their teaching practice. Possible considered judgments include equality, fairness, human dignity, respect for persons, freedom, among many others. Possible principles in reflective equilibrium with those convictions include the following principles:

• Principle of Security: The privacy and security of sensitive information should be ensured. • Principle of Non-discrimination: AIED design should avoid invidious discrimination based upon morally arbitrary characteristics of persons and groups; AIED technology design and development should train applications with comprehensive and diverse data to reduce bias. • Principle of Privacy and Confidentiality: Well-informed consent from users and the maintenance of the confidentiality of the users’ information should be guaranteed, regarding both the information users provide and collected information about them. • Principle of No Harm: Only AIED systems should be designed and deployed that protects users from harm, including academic and political repression. • Principle of Human-Centeredness: AIED technology should complement and enhance the development of human cognitive, intellectual, social, civic, economic, and cultural capabilities and well-being. • Principle of Honesty: The ethical use of AIED technology should be enforced regarding cheating, intellectual and academic fraud, and dishonesty. • Principle of Digital Equity: Fair and equitable access to and use of AIED technology should be ensured. • Principle of Academic Freedom: In all aspects of the ethical development and use of AIED technology the academic freedom of parties should be protected.