Skip to content
Theoria home
Log in

For teachers in upper secondary school

Using a current study in class: a plan for one lesson

New studies make good lessons: they are current, often contested and show how knowledge is produced. The hard part is choosing one students can follow and preparing it without losing an evening. This guide walks through one lesson, from choosing the study to discussing what remains open.

Updated on 20 October 2026

The question

How can I bring a current scientific study into one lesson so that students work with the source and not just with a summary?

Step by step

  1. Step 1

    Choose a study with a clear question

    Good candidates ask a question your students care about, have a design you can explain in two sentences and a freely accessible abstract, ideally the full text. Avoid studies whose main result only makes sense with advanced statistics.

    In Theoria: Browse new papers by category or search for a topic. Every paper links to its original.

  2. Step 2

    Check what the study can carry

    Before class, note the type of study, the sample, the comparison, the publication status and any published correction. Pick one limitation that students should discover themselves.

    In Theoria: AI Peer Review lists strengths and limitations of a paper as a starting point for your own check.

  3. Step 3

    Prepare the material at two depths

    Same source, different depth: a short version for everyone and more detail on the methods for students who want to go further. Keep the original passages that matter for the lesson at hand.

    In Theoria: Deep Insights explains a paper at a depth from 1 to 10. For teachers with Theoria Education, the Teacher Lesson Kit plans a lesson from a paper by grade band, duration and lesson focus, with a teacher PDF and a student handout. It needs enough full text of the paper.

  4. Step 4

    Run the lesson in four phases

    5 minutes: present the research question and let students predict the result. 15 minutes: in pairs, students find the sample, the comparison and the main result in the original. 15 minutes: students judge three statements as supported, contradicted or not decidable with these data, each with the passage as evidence. 10 minutes: discuss what we still do not know and what a follow-up study would need.

  5. Step 5

    Close with transparency

    Name which tools prepared the material and which statements you checked yourself. That way students see that AI is a tool for reading, not an authority.

Worked example

A field experiment on AI tutors in high school mathematics

In June 2025 Hamsa Bastani and colleagues published a field experiment in PNAS in which nearly a thousand high school mathematics students received access to AI tutors. It suits a lesson on AI and learning because the people studied are students themselves.

Design
Two tutors based on GPT-4: one that mimics a standard ChatGPT interface ("GPT Base") and one with prompts designed to safeguard learning ("GPT Tutor").
Result with access
Grades on practice problems improved by 48 percent with GPT Base and by 127 percent with GPT Tutor.
Result without access
When access was taken away, students who had used GPT Base did 17 percent worse than students who never had access. The safeguards in GPT Tutor largely mitigated this effect.
A statement to check
"Using AI makes students better at maths." The source supports this for practice with access and contradicts it for later work without access, at least for GPT Base.
Before you use it
PNAS published a correction to this article in August 2025. Read it before class to see whether it affects the points you want to use.

One study, two effects that point in different directions. That makes it a good case for separating what was measured from what is claimed.

Where AI reaches its limits

  • AI can propose a lesson structure, but it does not know your class. Check level, time and language yourself.
  • AI summaries and tasks can contain errors. Check every statement students are meant to verify against the original passage before the lesson.
  • One study is not a body of evidence. Present it as a current finding, not as the final answer.
  • Follow your school's rules on AI use and data protection before students use AI tools.

Try it with your class

The page for schools shows how teachers and students work with sources in Theoria. If you have questions about using it at your school, write to us.

Teachers with Theoria Education can also prepare a learning path and assign it to their students: in stations, students choose between plausible conclusions, justify their choice with evidence from real papers and judge statements as supported, contradicted or not decidable with the data.

Theoria for schoolsWrite to us

Sources

  1. Bastani, H., Bastani, O., Sungu, A., Ge, H., Kabakcı, Ö., & Mariman, R. (2025). Generative AI without guardrails can harm learning: Evidence from high school mathematics. Proceedings of the National Academy of Sciences, 122(26), e2422633122. https://doi.org/10.1073/pnas.2422633122Open access. Basis of the worked example.
  2. Correction for Bastani et al., Generative AI without guardrails can harm learning: Evidence from high school mathematics (2025). Proceedings of the National Academy of Sciences, 122(34), e2518204122. https://doi.org/10.1073/pnas.2518204122Published correction from August 2025; read it together with the article.

Publication details checked against Crossref and arXiv on 20 October 2026.

Other guides

All guides