Defining Desirable Results from a GenAI Tool: Hallucinations, Undesirable Results, and ISO/IEC AI Quality Standards

This podcast, based on the work of Eric E. Cohen, CPA, and published by ThinkTWENTY20, explores how we can better define and address the quality of outputs from Generative AI systems. Cohen challenges the overuse of the term "hallucination" to describe undesirable AI results and instead calls for a deeper understanding rooted in user experience, external influences, and system design. Drawing on recent developments from the International Organization for Standardization (ISO), including ISO/IEC TS 25058:2024, the discussion highlights how emerging standards can guide AI developers and users toward more trustworthy, useful, and ethically responsible AI systems. The podcast encourages a human-centric approach to AI—one that emphasizes trust, satisfaction, and societal benefit.

Defining Desirable Results from a GenAI Tool: Hallucinations, Undesirable Results, and ISO/IEC AI Quality Standards

Course curriculum

Part 1: Defining Desirable Results from a GenAI Tool: Hallucinations, Undesirable Results, and ISO/IEC AI Quality Standards

Part 2: Defining Desirable Results from a GenAI Tool: Hallucinations, Undesirable Results, and ISO/IEC AI Quality Standards

Part 3: Defining Desirable Results from a GenAI Tool: Hallucinations, Undesirable Results, and ISO/IEC AI Quality Standards

About this course

Course Overview

Discover your potential, starting today