For Academic Librarianss

20 Practical Ideas for Academic Librarians to Stay Cognitively Sovereign

Students are submitting literature reviews built entirely on AI summaries, never reading original papers themselves. Your role is to rebuild the skill of evaluating sources before trusting what AI claims about them.

These are suggestions. Take what fits, leave the rest.

Download printable PDF

Teaching Source Verification

Ask students to verify one AI citationbeginner

Have them find the actual paper and check if AI quote exists in it.

Show how AI creates plausible but false referencesbeginner

Compare Elicit results to your library catalogue to expose hallucinated journal names.

Teach DOI and ISBN lookup as verification stepbeginner

Train students to check every citation against CrossRef before including it in work.

Create assignment requiring source disagreement with AIintermediate

Ask students to find one paper AI missed or misrepresented in its summary.

Run workshop on evaluating Semantic Scholar resultsintermediate

Show when citation counts mislead and how to spot predatory papers AI recommends.

Demonstrate Connected Papers network reading strategyintermediate

Show how to read map visually instead of accepting AI summary of cluster meaning.

Build checklist for AI bibliography red flagsbeginner

List signs of hallucination: mismatched author names, impossible publication dates, fictional venues.

Teach how to read abstract before trusting AI synopsisbeginner

Spend five minutes on original abstract to catch AI misinterpretation of findings.

Create comparison between human and AI literature reviewintermediate

Show students where Perplexity missed key papers or overstated consensus.

Assign critical reading of AI-generated research summaryintermediate

Have students annotate where AI claim lacks support in cited sources.

Protecting Research Integrity

Establish library policy on AI tool transparencyintermediate

Require students to document which AI tools they used and for what tasks.

Create mandatory source access requirement for conclusionsbeginner

Students must read at least one full paper supporting each major claim.

Train students to interrogate AI about source uncertaintybeginner

Teach them to ask ChatGPT if it is confident in citations before using them.

Develop peer review protocol checking for AI hallucinationsintermediate

Train student reviewers to verify citations before approving papers for submission.

Build database of AI errors in your disciplineadvanced

Document specific hallucinations Elicit or Semantic Scholar make in your research area.

Teach when to reject AI research design suggestionsadvanced

Show students how AI methodology recommendations can violate your field's ethics standards.

Create student research log documenting AI decisionsintermediate

Have them record why they accepted or rejected each AI suggestion in literature review.

Establish office hours focused on source evaluationintermediate

Meet with researchers before they submit to verify their most critical citations.

Require abstract reading before citing AI-summarised papersbeginner

Make it a visible checkpoint in your library research appointment template.

Design assignment comparing AI versus archival researchundefined

Have students find sources using both tools to see what each method misses.

Five things worth remembering

Keep a shared document of AI errors found by your library users.
When students cite AI summaries, ask them to show you the original PDF.
Teach the five-minute rule: read the abstract before trusting any AI summary.
Train students to say 'I verified this source' instead of 'AI found this source'.
Build relationships with faculty so they notice when student work bypasses deep reading.

20 Practical Ideas for Academic Librarians to Stay Cognitively Sovereign

Teaching Source Verification

Protecting Research Integrity

Cognitive Sovereignty: How To Think For Yourself When AI Thinks For You