WEEK 2 GENERATIVE AI IN HIGH ED

Four "Good" advantages — name 2:

Enhancing the Learning Experience; Instructional Benefits (also: AI Benefits for Reading; AI Benefits for Writing)

Four "Bad" disadvantages — name 2:

Fabricated Literature Reviews; Misrepresentation of GenAI Output as Your Own Work (also: Plagiarism; Refinement and Fabrication of Fake Content)

Four "Ugly" negative externalities — name 2:

Composing Essays and Computer Programming (reduced cognitive engagement); Labor Arbitraging (also: Inflated Self-Efficacy; Cognitive Decline)

approaches to literature reviews — list 3:

Exploratory reviews; Integrative reviews; Meta-narrative reviews (also: Critical reviews; Meta-analysis; Meta-regression; Mixed research syntheses)

T/F: Citation fabrication in 18% of ChatGPT-generated literature reviews in one study.

True

Citation fabrication attributed to GenAI being fundamentally a text transformer rather than an information retrieval system.

False

Yeardon et al. showed GenAI is very effective at producing high-quality physics essays.

True

Westfall's research shows GenAI does poorly on the USMLE.

False

Bartz et al. v. Anthropic ruled that harvesting copyrighted materials for LLM training is justified under "fair use."

YES

T/F: No evidence that GenAI use in essay writing reduces cognitive load enough to interfere with cognitive engagement.

False

Evidence GenAI users show different patterns of cognition.

YES

Klopfer MIT study — which control group showed greatest understanding?

students who used only a search engine (100% passed)

Which group completed assignments fastest?

students who used ChatGPT

Term for using GenAI to reduce labor costs:

Labor arbitraging

T/F: Cognitive offloading to GenAI has NOT been shown to negatively correlate with critical thinking.

False

YES/NO: Can our experience with GenAI confirm the Dunning-Kruger Effect?

No