Home/Discussion Groups/Session 8: AI, Expertise, and the Future

Session 8: AI, Expertise, and the Future

This session uses credencing to discuss AI systems, expert judgment, overtrust, and future decision-making.

Concept Focus

AI makes credencing more important because it can produce fluent answers that feel authoritative. A user must ask: what evidence supports this output, how confident should I be, and what checks are needed before acting?

Expertise also requires credencing. Experts deserve more weight inside their domain, but expertise is not unlimited authority and does not remove uncertainty.

Historical Example

Earlier technologies such as printing, radio, television, and search engines changed how people formed beliefs. Each made some knowledge easier to share, but each also created new ways to spread error or propaganda.

AI continues that pattern. The issue is not only whether the tool is powerful, but how people assign confidence to what the tool produces.

Contemporary Issue

Students, doctors, lawyers, journalists, and managers now face AI-assisted work. Groups can discuss where AI is useful, where verification is essential, and when human overtrust becomes the main danger.

Discussion Questions

Questions with expandable guidance

Open each question for facilitator notes, examples, scenarios, and feedback. The aim is careful discussion at a senior high school reading level.

Question 1Why can a fluent AI answer feel more reliable than it is?+

GuidanceFluency is not the same as evidence. Ask what sources, checks, uncertainty, and domain knowledge support the output.

Historical ExampleEarlier technologies such as radio, television, and search engines also made information feel more authoritative when it was easier to access.

FeedbackThe key distinction is presentation quality versus evidential support.

Question 2When should people defer to experts, and when should they ask for more explanation?+

GuidanceExpertise matters most inside a domain and when experts show methods, uncertainty, and track records. Deference should not become blind trust.

Hypothetical ScenarioA doctor, an AI chatbot, and a social media influencer give different health advice. What questions should a patient ask before assigning confidence?

FeedbackStrong answers weigh domain expertise, evidence, incentives, and verification.

Question 3What future institutions will need better confidence governance?+

GuidanceDiscuss schools, courts, hospitals, AI labs, journalism, government agencies, and companies.

Contemporary IssueAI-generated content, deepfakes, automated decisions, and model-based forecasts all require clearer confidence labels and accountability.

FeedbackGood answers focus on systems that make confidence visible and correctable.