The Wizard Problem, Trust, Transparency and Teaching in the Age of AI | Ep. 77

In this episode of ChatEDU (The Wizard Problem, Trust, Transparency and Teaching in the Age of AI), Matt and Liz start with a deepfake dress-up, a Saturday Night Fever–style keynote pose, and Google’s viral “Nano Banana”. They spotlight new Gemini features and predict moves in email, Outlook, and Google’s AI strategy. Then it’s a rapid rundown, a case for why AI won’t rival human intuition, and a dive into “wizard” agents.Story #1: Rundown RoundupAI Humanizers Flop: Lifehacker tests Paraphraser.io and others; most fail detectors. StealthWriter helped somewhat, but inconsistently.Turnitin Strikes Back: New detectors catch AI writing and text rewritten by humanizers.Gallup AI Survey: 98% of Americans know about AI, but only 8% feel “very knowledgeable.” Trust rises with use, but job fears persist.Pew on AI: 53% think AI will hurt creativity; most want more control and worry about telling AI from humans.OpenAI User Study: 700M use ChatGPT weekly; 10% of use is education-related. Younger users dominate; gender gap closed.Anthropic Trends: Workplace AI adoption doubled since 2023; instructional material creation up 6x. Utah and D.C. lead.Teen Plan: OpenAI adds age prediction, defaulting to under-18 mode when unsure. Parents get controls and alerts.Altman Speaks: In a Tucker Carlson interview, OpenAI’s CEO says his fear isn’t doomsday but everyday misuse.AI Textbooks: Google’s “Learn Your Way” tool personalizes reading and assessment, outperforming eBooks in early trials.Story #2: Why AI Will Never Run the WorldMatt and Liz unpack Angus Fletcher’s Primal Intelligence, which argues AI can’t replicate human “story thinking.” AI is strong in logic and prediction but weak in intuition, imagination, emotion, and common sense. Liz frames literature as training for flexible thinking; Matt applies Fletcher’s ideas to front office triage. They explore how this divide affects education, leadership, and innovation.Story #3 (Beneath the Surface): The Wizard Behind the CurtainEthan Mollick says AI acts like “wizards,” producing polished work with little transparency. Liz tests this by giving Claude and it outputs a McKinsey-style presentation with insights and formatting. But in Massachusetts, an AI glitch misgraded 1,400 student essays on the MCAS test. The lesson: with wizardry comes responsibility.Bright Byte: Save the Orange!Coca-Cola and MIT use generative AI to fight citrus greening, which threatens orange juice worldwide. By simulating disease spread and testing interventions virtually, they show how AI can tackle agricultural crises.Announcements and SponsorsCheck out the new Student AI Course for middle and high school. Email Matt and Liz at [e-mail] Fall Micro-Credential is open. Join us for the educator and school leader course starting in October - skills21.org/ai/microThe National Center for Next Generation Manufacturing. Explore their October 15–17 Summit in Rochester Hills, MI. www.nextgenmfg.orgEDIAThe AI-powered platform helping schools reduce absences and boost achievement. edia.app/contactLinks and ReferencesGemini Gem Sharinghttps://tinyurl.com/k4ah3689Google’s Learn Your Way⁠https://tinyurl.com/3v6rd9td⁠Lifehacker: AI Humanizershttps://tinyurl.com/4kv842bsTurnitin Detection Updatehttps://tinyurl.com/3u2pxzz7Gallup AI Studyhttps://tinyurl.com/kesa6xehPew AI Reporthttps://tinyurl.com/mryp2dvcOpenAI User Studyhttps://tinyurl.com/b82bs99kAnthropic Economic Index⁠https://tinyurl.com/37stk67h⁠OpenAI Teen Safety Updateshttps://tinyurl.com/5n7j5sfz⁠https://tinyurl.com/2cz87pvcSam Altman Interviewhttps://tinyurl.com/4sv68ps9Angus Fletcher’s Primal Intelligencehttps://tinyurl.com/2d9zujj6Ethan Mollick: On Working with Wizardshttps://tinyurl.com/4n59dhsnSkills21 Prompt Review Toolhttps://www.skills21.org/promptsCommon Sense Media AI Toolkithttps://tinyurl.com/5xet3e4tBright Byte: Save the Orangehttps://tinyurl.com/4f6mxpm7