Psychology of AI 5: Agentic AI Autonomy, Control and Human Responsibility
About This Course
This two-day standalone course examines what happens when Artificial Intelligence (AI) systems move beyond generating answers to planning tasks, calling tools, triggering workflows, communicating with other systems and acting on behalf of humans. Participants will explore the psychological, ethical and governance implications of agentic AI, with a particular focus on autonomy, control, human agency, assurance, interface design, indicators and intentionality.
Participants will examine practical issues such as autonomy boundaries, agency controls, assurance mechanisms, human-reliance indicators, interface risks, AI-to-AI interaction, escalation design and accountability. The course uses an applied ethical and operational lens covering whether the system can act independently, what controls and overrides exist, how safety and trustworthiness are assured, what metrics matter beyond productivity, how humans interact with the system, and who determines its purpose and objectives.
What You'll Learn
- Explain the shift from Artificial Intelligence (AI) assistance to agentic AI, including how AI systems can plan, call tools, trigger workflows and act on behalf of users.
- Analyse how agentic AI affects human agency, control, awareness, accountability and responsibility.
- Identify key agentic AI risks, including excessive autonomy, weak oversight, agentic overload, AI-to-AI opacity, poor escalation design and unclear accountability.
- Define appropriate autonomy boundaries, control points, human override mechanisms and escalation triggers for agentic AI use cases.
- Apply governance and assurance principles to assess the safety, reliability, trustworthiness and accountability of AI agents.
- Design meaningful indicators and interfaces that help humans monitor, understand, intervene in and govern agentic AI behaviour.
- Develop a practical Agentic AI Responsibility Map or risk assessment covering autonomy, human control, assurance, indicators, interfaces, intentionality and accountability.
Entry Requirements
Participants should preferably hold a degree or diploma and have at least three years of relevant work experience.