Use the stakeholder-by-value norm matrix as actionable rules for implementing ethical AIED¶

Strategy · All strategies
Evidence · no claims cited

Description¶

The review proposes its combined matrix as a practical tool: ethical norms are treated as "actionable rules that could be used to achieve the desired behaviours described by the EVs". Stakeholders can look up their set (developers, educational institutes, end users, regulators) and find the norms the literature assigns for each main ethical value, e.g. educational institutes carrying most human-oversight norms and regulators most goodwill norms. The grouping per stakeholder set helps different stakeholders ensure AI used in education is ethical, and the grouping per value supports implementing AIED in compliance with specific values.

Design Implications¶

Context¶

Requirements¶

• Stakeholders must first identify which of the four stakeholder sets they belong to; norms must be matched to the specific main ethical value being implemented

Constraints¶

• Norms derive from only 15 articles providing norms, with uneven coverage across values and stakeholder sets; end-user norms are sparse

Target Learners¶

• educational leaders, AIED developers, policymakers, teachers, and students

Target Learning Goals¶

• implementing ethical values in the design, deployment, and governance of AIED systems

Claims¶

Related Strategies¶

• Norm Setting

Examples¶

-

Key Sources¶

• Agarwal, B., Urlings, C., van Lankveld, G., & Klemke, R. (2022). Identifying the Ethical Values and Norms of Artificial Intelligence in Education: a Systematic Literature Review. Open University of the Netherlands. https://orcid.org/0000-0002-7347-8465