Summary

Artificial intelligence divides students into winners and losers. A workshop in Basel shows: many teenagers use AI tools like Chat-GPT without critically reflecting on the technology. While some students achieve better grades with it, others refrain from delegating their thinking to machines and thus lose central learning competencies. Experts warn of a growing digital divide that exacerbates existing educational inequalities.

People

Topics

  • Artificial intelligence in the classroom
  • Digital competency of teenagers
  • Educational inequality
  • Critical thinking
  • AI tutors

Clarus Lead

Students use AI tools like Chat-GPT extensively but cannot handle them competently. A Basel school workshop documents how some teenagers are already falling behind, while others strategically use the technology and benefit from it. Researcher Marc Eyer warns: «AI leads to a new digital divide.» Society is splitting into users with media literacy and those without – with direct consequences for school performance and future opportunities. Schools must respond urgently as teachers often ignore reality.

Detailed Summary

At St. Alban Secondary School in Basel, reality shows: 16 students are supposed to learn to use artificial intelligence appropriately. Yet at the very first example – an AI that generates salmon fillet instead of swimming salmon – many fail. The educator must present the solution herself. The problem is not a lack of access to technology, but rather a lack of media literacy. One student took content from Chat-GPT for a book presentation without realizing that the AI had invented information – and received a poor grade for it.

In contrast: a Zurich high school student read none of his German final exam books, had Chat-GPT explain the material to him, and passed with top marks. The same tool, two completely different results. Marc Eyer from the Pedagogical University of Bern analyzes the phenomenon as a «digital divide». Similar to earlier technologies, two groups emerge: those who have access and competency, and those who lack both. Students with high media literacy learn faster; others risk falling further behind.

Eyer criticizes that some teachers deny that AI plays a role in their teaching. He calls for strategic professional development. At the same time, research warns: learning must be demanding. Those who do not challenge their brains forget more easily. His team is therefore developing AI tutors that do not simply provide answers but guide students toward independent thinking. Today, however, students mostly use free standard tools. A 15-year-old admits openly: he has «no desire» to read; Chat-GPT spares him the work.

Key Takeaways

  • Two-way split: Some students benefit from AI competency; others lose through lack of reflection and critical thinking practice
  • Hallucinations and uncritical use: Many teenagers use AI output without verification and do not notice when information is fabricated
  • New inequality: The existing concept of the «digital divide» is intensifying; AI amplifies social educational inequalities
  • Systemic failure: Schools do not rely on specialized AI tutors that encourage thinking; standard tools enable shortcuts without learning effect
  • Professional development shortage: Teachers are often not equipped for AI integration; some completely ignore the phenomenon

Critical Questions

  1. Evidence/Data Quality: On what representative data is Eyer's statement about the «new digital divide» based? Was a baseline study conducted to measure the effect?

  2. Conflicts of Interest: Does Eyer's team develop AI tutors commercially? If so, could self-interest influence criticism of standard tools?

  3. Causality: Are poor grades attributed to AI misuse, or did such students already have lower motivation/abilities? Was the control for prior knowledge sufficient?

  4. Alternatives: Instead of banning AI – how do other countries handle specialized AI education? Are there best practice examples?

  5. Feasibility: What concrete professional development for teachers is practical? Timeline? Costs? Who finances it?

  6. Side Effects: Could an AI ban in certain subjects (as the report suggests) lead to even larger performance gaps if some students continue learning with AI privately?

  7. Definition of Media Literacy: How is it measured whether a student handles AI «competently»? Are there clear criteria or does this remain subjective?

  8. Generational Effect: Is the «lost generation» really lost, or do these students learn more independently through error experiences later?

Sources

Primary Source: Auch Jugendliche werden abgehängt: KI öffnet eine neue digitale Kluft in der Bildung – NZZ, 15.02.2026

Supplementary Sources (mentioned in text):

  1. Podcast «NZZ Quantum Leap»: AI tutors and critical thinking (19.12.2025)
  2. Gioia da Silva & Leonid Leiva Ariosa: «Those who constantly ask Chat-GPT practice critical thinking less» (06.07.2025)
  3. Eduard Kaeser: Guest commentary on Chat-GPT and writing (08.12.2025)

Verification Status: ✓ 15.02.2026

This text was created with the assistance of an AI model. Editorial responsibility: clarus.news | Fact-check: 15.02.2026