Podcast: Play in new window | Download
What if the words we use to describe AI are misleading us—and even shaping how we regulate, trust, and collaborate with technology itself?
In this revelatory episode of the Tech Humanist Show, host Kate O’Neill sits down with Anna Mills and Nate Angell, co-authors of a provocative new paper arguing that the term “AI hallucination” may be doing more harm than good.
Together, they explore why “mirage” is a much better metaphor for those times AI gets things wrong—one that empowers human users and encourages critical thinking, rather than mystifying the technology or making it seem almost sentient.
Topics Covered:
Background and Motivation for the AI “Mirage” Paper
Origins and Issues with the Term “AI Hallucination”
The Case for “AI Mirage” as a Better Metaphor
Deep Dive into the Mirage Metaphor
Power Dynamics, Language, and Metaphors in AI Discourse
Implications for Educators, Technologists, and Policymakers
The Publication Process and Interdisciplinary Collaboration
Adoption, Use, and Evolution of AI Terminology
Human Agency, Hope, and Creativity in AI’s Future
Connect with Anna Mills and Nate Angell:
LinkedIn: Anna Mills + Nate Angell
Paper: Are We Tripping? The Mirage of AI Hallucinations
📋 Episode Chapters
00:00:05 – Welcome and introductions: Why language matters in AI
00:02:24 – Anna and Nate introduce their work and the AI Mirage paper
00:05:07 – Where did “hallucination” come from—and why is it problematic?
00:08:49 – How metaphors shape policy, public understanding, and AI design
00:10:18 – The problematic implications of “hallucination”: Stigma and technical misconceptions
00:11:48 – Unpacking the “mirage” metaphor: Origins, process, and collaborative input
00:18:44 – Why “mirage” works: Human judgment, desire, and skepticism
00:22:02 – The complexity of AI “mirages” versus human interpretation
00:25:49 – Language, power dynamics, and shaping critical AI literacy
00:32:05 – Advice for educators and technologists: Building agency and literacy
00:33:08 – What would change if “mirage” replaced “hallucination”?
00:35:21 – Interdisciplinary collaboration and the importance of social discourse
00:38:33 – Practical tips for adopting better language about AI
00:39:23 – Staying hopeful and humanistic in a rapidly evolving tech landscape
00:41:30 – When AI “mirages” are actually “rainbows”: Embracing creative delight
00:42:37 – Closing thanks and reflections
