Episode Highlights
- From AI Output to Action (11:52)
Introducing the central challenge of the session: how to decide when an AI-generated output is trustworthy enough to act upon.
- When AI Makes Bias Harder to Spot (24:15)
Exploring how AI can reinforce existing biases while making them less visible through persuasive and highly confident responses.
- A Safety Incident That Changed the Story (27:18)
Using a practical safety scenario to show how recent events can skew AI recommendations when broader context is missing.
- Making Bias Visible Before It Shapes Decisions (32:30)
Sharing a four-step framework for surfacing assumptions, testing conclusions and validating evidence in AI-supported work.
- The Ethical Questions AI Can’t Answer for You (34:39)
Moving beyond technical capability to examine the human responsibilities of accountability, transparency and fairness.
- The Risks of Screening People with AI (38:47)
Using an apprenticeship example to highlight the ethical risks and unintended consequences of AI-supported decision making.
- Eight Tests for Better AI Decisions (43:29)
Introducing practical decision tests designed to challenge assumptions and encourage more responsible use of AI.
- Earning Confidence in AI Recommendations (48:08)
Examining why trust comes from transparency, feedback and local knowledge rather than simply accepting AI outputs as correct.
- The PAUSE Model for Responsible AI Use (51:22)
Bringing the session together with a practical framework for applying critical thinking to AI-supported decisions.
- Why Critical Thinking Matters More Than Ever (56:43)
In the Extra Slice Q&A, discussing why the ability to question, evaluate and challenge information is becoming increasingly important in the age of AI.
