Back

Agentic AI in Practice

91-729

Units: 6

Description

Agentic Artificial Intelligence (Agentic AI) refers to systems composed of multiple autonomous agents powered by large language models that analyze problems, retrieve knowledge, reason over evidence, and generate policy recommendations or operational decisions. These systems are increasingly influencing how organizations and public institutions make decisions.

This course introduces students to Agentic AI through a hands-on policy laboratory approach. Students interact directly with a deployed Agentic AI platform hosted on AWS. Using this platform students design policy scenarios, run simulations, observe AI reasoning traces, evaluate system outputs, and identify risks such as hallucination, bias, unsupported evidence, and incomplete reasoning.

The course emphasizes learning through experimentation. Students test the behavior of agentic systems in practice and develop governance frameworks for their responsible use in public policy contexts.

Students initially apply the system to public health policy scenarios and later transfer their analytical methods to public housing and criminal justice policy domains.

Learning Outcomes

By the end of this course, students will be able to:

1. Understand how agentic AI systems operate as decision systems rather than conversational tools.
2. Translate a policy problem into structured inputs suitable for AI analysis.
3. Run simulations using an agentic AI platform and interpret system reasoning.
4. Identify risks such as hallucination, bias, unsupported evidence generation, and incomplete reasoning.
5. Evaluate governance and ethical implications of AI decision systems.
6. Transfer analytical methods across public health, housing, and criminal justice policy domains.
7. Design guardrails for responsible AI deployment.

Prerequisites Description

• None