A recent experiment in middle school classrooms demonstrated a simple but powerful lesson: today’s AI hype doesn’t necessarily equate to genuine intelligence. When students interacted with ELIZA, a 1960s chatbot designed to mimic a therapist by simply reflecting user statements back as questions, their frustration was immediate. One student called the bot “gaslighting,” while another bluntly stated it was unhelpful.
This wasn’t about rejecting AI outright. Rather, the exercise, conducted as part of EdSurge Research’s investigation into AI literacy in schools, revealed a critical gap between perceived intelligence and actual functionality. The core takeaway: understanding how AI works is far more valuable than merely using it.
The Lesson: Deconstructing the Illusion
The teacher intentionally chose ELIZA, a deliberately rudimentary program, to force students to confront the limitations of early AI. The bot’s endless “tell me more” prompts and robotic deflection (“We were discussing you, not me”) quickly exposed its superficiality.
This frustration wasn’t the bug, but the feature. The teacher aimed to cultivate what learning scientists call productive struggle — the discomfort that arises when students are challenged to think critically. By building their own chatbots using MIT App Inventor, students were forced to grapple with the fundamental mechanics of AI. They discovered that without extensive training data, even a simple chatbot remains fundamentally incapable of true understanding.
The point wasn’t about making students better coders, but about teaching them how to decompose complex systems into manageable parts. This process builds frustration tolerance, a crucial skill for tackling difficult cognitive tasks, and cultivates computational thinking – the ability to break down problems into logical steps.
Why Computational Thinking Matters More Than Coding
As Columbia University’s Jeannette Wing argues, “Computers are dull and boring; humans are clever and imaginative.” The focus should be on developing the uniquely human skills that AI can’t replicate, like critical thinking, empathy, and problem-solving. The AI boom reinforces this. Coding skills may become obsolete, but the ability to reason about systems, interrogate outputs, and distinguish between superficial fluency and actual understanding will remain essential.
The experiment highlights a disturbing but predictable trend: students understand that AI tools like ChatGPT have flaws (“It can sometimes give you the wrong answer”). Yet, they continue to rely on them because they are perceived as useful and efficient. As one student put it, “I just want AI to help me get through school.”
The Bigger Picture: Skills Over Tools
The lesson didn’t eliminate students’ dependence on AI, but it did demystify the technology. They learned that chatbots operate through prediction, not genuine intelligence, and that trust in these tools is often driven by social signaling rather than understanding.
Educators must prioritize teaching the thinking behind AI, not just the tools themselves. This approach provides students with a durable skill set that will outlast any specific technology. The ability to ask better questions, evaluate outputs critically, and recognize the inherent limitations of AI will be far more valuable in the long run than simple tool proficiency.
The lesson underscores the importance of educators’ discretion. In a climate where schools are pressured to either embrace or reject AI, guided instruction and ethical considerations are paramount. Understanding how chatbots work is the first step toward responsible AI usage – a skill that will matter long after today’s tools become obsolete.
