TEACHER VOICE: AI is an addictive drug that must be researched, studied and confined
Source: The Hechinger Report
Author: Mike Kentz
Original source: https://hechingerreport.org/teacher-voice-ai-is-an-addictive-drug-that-must-be-researched-studied-and-confined/
Related source: What “Just Say No” Got Wrong About AI
Published: 2026-06-01
Source type: Essay
Private backup: the full source text is archived in the private repository at archives/articles/hechingerreport-org-teacher-voice-ai-is-an-addictive-drug-that-must-be-researched-studied-and-confined.source.md. The archive is not published on the public Quartz site.
Summary
Mike Kentz argues that schools should treat AI less like a tool students can simply be told to avoid and more like a powerful substance that requires research, containment, education, and monitored exposure. He uses a classroom character-bot interrogation activity as evidence that structured encounters can help students become less captivated by AI, not more dependent on it.
The essay’s central contribution is the vaccine metaphor: AI literacy should help students build “antibodies” through bounded, reflective experience. Kentz does not minimize AI’s risks; he argues that because students already encounter AI outside school, the stronger response is proactive AI literacy that teaches students to interrogate systems, understand harms, and develop discernment.
Pull quotes
Structured encounter
Those students didn’t need to be told AI was addictive. They didn’t need a policy or a warning. They came to their own conclusions because they had been given a structured encounter with the technology — one that required them to interrogate it rather than consume it.
— Mike Kentz, TEACHER VOICE: AI is an addictive drug that must be researched, studied and confined
AI literacy as antibodies
I believe that is what AI literacy actually produces. Not dependence — antibodies.
— Mike Kentz, TEACHER VOICE: AI is an addictive drug that must be researched, studied and confined
Monitored exposure
Physiological resistance is developed through exposure. Not through uncontained exposure, but via monitored exposure.
— Mike Kentz, TEACHER VOICE: AI is an addictive drug that must be researched, studied and confined
Drug education works
With respect to narcotics, we know that “Just Say No to Drugs” doesn’t work. Drug education works. Drug literacy works.
— Mike Kentz, TEACHER VOICE: AI is an addictive drug that must be researched, studied and confined
Big ideas
- AI simulations need clear boundaries for learning
- AI literacy requires different kinds of AI interaction
- Schools should start with learning values before choosing AI tools
- District AI work is a long-term redesign project
Claims
- AI literacy builds resistance through structured exposure
- Punitive AI bans can drive student use underground
- AI literacy should teach students what to do with AI, not just what to think about it
- Students need boundaries for when to use AI and when to step back
Key evidence and examples
- Kentz describes a classroom assignment where students interrogated fictional character bots and later reported little interest in repeating the experience, which he reads as evidence that structured critique can reduce uncritical fascination.
- He argues that AI literacy is not capitulation to AI adoption; it can be a way to build resistance, discernment, and awareness of harms.
- The essay frames abstinence-only AI policy as inadequate because students already encounter AI in uncontained settings.
- Kentz calls for research, documentation, dosage, and monitored exposure rather than either blanket embrace or simple prohibition.
Education relevance
High relevance for K-12 AI literacy, school AI policy, student AI use, bounded classroom experimentation, and teaching students to interrogate AI systems rather than merely consume them.
Durability note
Durability: High. The drug/vaccine metaphor is provocative and contestable, but the underlying policy distinction between uncontained exposure, abstinence-only warnings, and structured monitored AI literacy is likely to recur.