Freshman Research, Copy-Paste Citations, and AI Summaries
You are the media specialist at a high school of 1,900 students. A 9th-grade ELA teacher and a biology teacher have assigned a joint research paper on a local environmental issue. There are 140 students. The teachers have booked the library for one 45-minute session, at the end of which students must submit a draft. In the drafts you see paragraphs that read like chatbot output, citations that list ChatGPT or simply Google, and links to content farms. The district has purchased three databases, but usage last year was 12% of licensed capacity. The district's AI guidance is a one-page acceptable-use document that says students must cite AI tools when used, but teachers say they do not know how to enforce it. The teachers want you to show them the databases in the single session. The principal wants a plan that improves the papers without delaying the unit.
1. Redesign the library's role in the unit. What instructional sequence would you propose, and what would you do in the first session?
Read the worked answer
I would start with backward design from the teachers' rubric: claim, evidence, source quality, citation, and disclosure of AI use. One 45-minute database demonstration cannot produce those outcomes, so I would propose a short co-taught sequence rather than a single visit.
Proposed sequence: Session 1 in the classroom or library, focused on narrowing the topic and generating searchable questions. Session 2 in the library, with stations for database searching, lateral reading, and citation capture. Session 3 back in the classroom, for revision and an AI-use disclosure conference. The teachers would assess the final paper; I would co-assess source quality and disclosure.
In the first session I would not begin with a database tour. I would show two anonymized draft excerpts, one with a content-farm source and one with a database source, and ask students to identify what a reader could verify. Then I would model one search from a student's topic, capture the citation, and have each student produce one usable source and one question before leaving. That gives the teachers formative evidence instead of a passive demonstration. I would also offer a short asynchronous guide so the teachers are not dependent on a single library period.
2. How would you teach source evaluation and AI disclosure so that the teachers can assess both?
Read the worked answer
I would make both skills visible and concrete.
Source evaluation: Teach lateral reading rather than a long checklist. Students would leave the source, open a new tab, and investigate the author, publisher, funding, and how other credible sources describe the issue. They would compare a content-farm page with a database article and a government or university source. The exit ticket would be a three-sentence source note: who is responsible, what evidence is offered, and why it is adequate for this claim.
AI disclosure: Students would complete a short AI-use log: tool used, prompt, what output was used, and how it was verified. The rule would be that AI can help generate questions or summarize background, but it cannot be cited as an author or used as the only source for a factual claim. Students would trace every factual claim to a verifiable source. The teachers can assess the log and the source notes with a simple rubric.
I would align this with the district's one-page guidance and ask the teachers to state the consequence for undisclosed use consistently. The goal is not to ban AI but to make the process auditable.
3. What evidence would you collect to show the principal that the redesign improved student work without overclaiming?
Read the worked answer
I would use a pre/post design with multiple measures, not a single satisfaction survey.
Measures: a common rubric applied to the first draft and the revised paper; the percentage of citations that point to credible, verifiable sources; the percentage of papers with a complete AI-use log; database sessions and source downloads during the unit; and teacher reports of time saved or problems encountered. I would also collect a short student exit ticket on whether they could explain how they checked a source.
Comparison: I would compare the same students' drafts and revisions, and, if available, the previous year's papers scored with the same rubric. I would not claim that the library session alone caused improvement, because the teachers also changed assignments and feedback. I would report the changes alongside what else changed.
Reporting: A one-page summary with three numbers, two student work excerpts, and one limitation would be more credible than a long promotional report. The principal needs a decision: continue the co-taught model, fund more database training, or adjust the schedule. I would recommend continuing only if the source-quality and disclosure measures improve.