Have students develop and train AI agents and simulations in the domains they are learning, with validation and scaffolding¶

Strategy · All strategies
Evidence · no claims cited

Description¶

In its scenario on developing professional practices, the article proposes coaching students to develop their own customizable AI agents for exploring a content area, or leveraging AI coding and rendering capacities to represent a domain. It reports that "Recently we've had students (who have no coding experience) designing computer simulations to represent basic to advance physics phenomena", requiring students to validate AI-based approaches for outcomes and scientific accuracy while learning the strengths and limitations of the technologies.

Design Implications¶

Context¶

Requirements¶

• Careful guidance and scaffolding is necessary, per the article, and students must validate AI-based approaches in terms of outcomes and scientific accuracy

Constraints¶

• The article presents this as a scenario with an early example rather than a systematically evaluated intervention

Target Learners¶

• undergraduate physics students, including those with no coding experience

Target Learning Goals¶

• communication, discernment, metacognition, trust, and understanding of the strengths and limitations of AI technologies

Related Strategies¶

• Extend simulated teaching practice to student assessment and teacher screening uses
• Provide differentiated scaffolding matched to teachers' behavioral-cognitive profiles when supporting multi-agent AI workflow design

Examples¶

-

Key Sources¶

• Finkelstein, N. (2026). A principled way to think about AI in education: guidance for educators and policy makers on action based on goals, models of human learning, and use of technologies. Draft manuscript, University of Colorado, Boulder. https://arxiv.org/abs/2510.01467. doi:10.48550/arxiv.2510.01467