by Sara Wink

Today’s college students must develop strong critical-thinking skills to navigate the seas of online information. Such skills are vital, as artificial intelligence (AI) tools and online misinformation tempt students with superficial assessments and automated summaries, diminishing critical thinking. What if we encourage students to move away from traditional methods, such as reliance on library databases and the CRAAP test, toward a framework that promotes lateral evaluation and wariness of AI-generated hallucinations? The CREDIBLE framework may be just the guide that aids students in assessing the unknown for what is truly reliable and trustworthy within the broader scholarly conversation.

Disney’s Aladdin is full of brilliant quips and songs, but one line has been on my mind a lot. It’s the moment Aladdin reaches out for Jasmine’s hand and asks, “Do you trust me?” (Musker & Clements, 1992).

Every time we go online, we are essentially offered hand after virtual hand, asking us to trust. Surely, this website is worthy. Surely, this speaker is an expert. Surely, this newspaper has worthwhile information.

But do they?

In this age of the hallucinated, the fabricated, and the outdated, it is more important than ever that we enable our students to critically assess research for use in the classroom and beyond.

The Struggle with Research Credibility

“If it’s .gov, that makes it okay.” “If it’s from the library, that makes it okay.” “If it’s from Wikipedia, that makes it NOT okay.”

We’ve heard these statements from students. We’ve likely said them ourselves, too. But as I have said to my students through the years, we should always have more than one reason to trust a source. Making a decision based on a domain extension is not much better than trusting a site because of how it is designed, or where it popped up in the search results.

In fact, all our pushing for library usage may encourage simple box-checking (Peer-reviewed? Found in a database?) rather than inspecting the logic and evidence in the information itself (Warner & Russo, 2026).

One can understand why students use box-checking, though, when we consider the density of the research we want them to study. Academic articles are full of specialized jargon and oodles of data, which can be tough for anyone to critically digest, let alone undergraduates (Warner & Russo, 2026). These students struggle to evaluate both traditional and digital sources. Even when the boxes are checked, they find it difficult to explain why those factors make a source trustworthy in the first place (van der Eem et al., 2025).

To further tangle matters, AI tools such as Grammarly, ChatGPT, Gemini, and others promise to simplify and explain these dense articles, removing any challenge for the student that would require mental struggle—the kind of struggle that can lead to frustration, sure, but growth, too. According to the College Board (2026), “More than 84% [of college faculty] agree that AI reduces students’ critical thinking, originality, and deep engagement with course material, while 88% are concerned about overreliance on automation” (para. 9). Such automation does more than ease cognitive load: it encourages students to rely on AI summaries without questioning what AI generates, removing the critical thinking step altogether. Traga Philippakos (2026) and Haider & Habiba (2025) documented that students struggle to critique information beyond the surface level, which is why undergraduates often concentrate on one or two criteria for assessing research.

The cognitive load is heavy indeed for students new to academic research, but tossing students into the deep end will not help them “discover” their new cognitive strengths in assessing credibility (Warner & Russo, 2026). Rather, it makes that shallow end of assessment with AI prompts all the more tempting.

Redefining the “Trustworthy” Test

The first time Aladdin asks Jasmine to trust him, they had worked together to escape an apple seller’s blade and patrolling guards. Jasmine had watched Aladdin in action and felt his company was worth keeping. The second time, she is NOT impressed with fancy-pants Prince Ali; the only reason she agrees to fly with Ali is that his question and mannerisms remind her of the boy from the street (Musker & Clements, 1992).

We have these moments as well, as researchers. A blog written by someone who used to work at a news station we watched surely must carry some weight. A publisher that was held in high regard in our own undergrad days surely must carry some weight. A website that looks just like the news publication we once grabbed from the newsstand, surely…you get the idea.

Educators must help students see that to be trustworthy communicators, they must evaluate what makes their own research trustworthy. How is the author an expert in what they wrote? How did those writers gather the research they shared (UC Irvine Libraries, 2026; van der Eem, et al., 2025)? How does the research of one source connect with the information shared in others (Haider & Habiba, 2025)?

For years, educators have taught the CRAAP test to help students see the importance of Currency, Relevance, Authority, Accuracy, and Purpose in their potential sources. These are all worthwhile elements to assess within information, but CRAAP predates AI-generated content with its authoritative tone and misattributed or hallucinated research (Traga Philippakos, 2026). CRAAP also encourages critiquing a source as an individual entity rather than reading laterally to see if other sources verify the original information (Haider & Habiba, 2025; Traga Philippakos, 2026).

Perhaps it’s time we leave the CRAAP in the gutter, and fly off with CREDIBLE. CREDIBLE encourages the very lateral reading that guides students to see a single source’s place in the bigger academic conversation, and that one publication can have ripple effects across many others.

Traga Philippakos (2026) breaks down this new assessment test in “CREDIBLE: A Framework for Critical Source Evaluation—From Information Consumers to Critical Evaluators:

C – Credibility: Who is behind the source, and is their process transparent? Can the author(s) be held accountable for the information they present?

R – Reliability: Does the source’s publisher have a documented track record of accuracy and consistency over time?

E – Evidence: Are claims supported by verifiable data, and do cited references actually exist? Where do these cited references take readers?

D – Date: Is the information current or does it reflect recent scholarship in the field     ?

I – Intent: Is the primary goal to inform, persuade, market, or entertain?

B – Bias: Are underlying perspectives, commercial ties, or omissions acknowledged? Are alternative viewpoints addressed?

L – Logic: Are arguments structurally sound, or do they rely on emotional appeals and generalizations?

E – Expertise: Do author/publisher qualifications match the specific research topic? (Traga Philippakos, 2026).

CREDIBLE’s test requires a far more layered assessment than CRAAP. Its questions work well not only for printed research but also for digital research. Plus, because AI can generate false or misattributed information, CREDIBLE motivates students to fly beyond basic summaries to view the source inches above the ground and up to heights within the broader scholarly conversation.

Share the Magic of CREDIBLE with Your Students

Just as Aladdin could not fly without the carpet, our students cannot magically transform their own critiquing skills without initial guidance (van der Eem et al., 2025). Repeated modeling of the process in a synchronous environment will provide students with the scaffolding knowledge they need to complete this test on their own. Consider working with non-academic articles first so students are more comfortable studying the information on their own (McGrew, 2020; Warner & Russo, 2026). Such modeling also provides opportunities to compare AI-generated output with credible research to help students see that glitter will not always mean gold.

References

College Board. (2026, Feb. 25). New College Board research: Faculty express near-universal concern that student AI use undermines original writing and critical thinking. Newsroom. https://newsroom.collegeboard.org/new-college-board-research-faculty-express-near-universal-concern-student-ai-use-undermines

Google. (2026, September 10). WACademic 2026 credibility research notebook [Generative AI chat]. Gemini. https://notebook.google.com/notebook/6501c7ee-07a2-437c-8de7-6148b42a6b3b/preview

Haider, N., & Habiba, U. (2025, June 13). From query to knowledge: Evaluating university students’ online searching proficiency and its connection to critical thinking and research skills. European Journal of Education, 60: e70144. https://doi.org/10.1111/ejed.70144

Musker, J., & Clements, R. (1992). Aladdin [Film]. Disney.

Traga Philippakos, Z. A. (2026). CREDIBLE: A framework for critical source evaluation—From information consumers to critical evaluators. AI in Education, 2(1), 3. https://doi.org/10.3390/aieduc2010003

UC Irvine Libraries. (2026, July 9). AI and information literacy. https://guides.lib.uci.edu/gen-ai/info-literacy

van der Eem, M., van Drie, J., Brand‐Gruwel, S., & van Boxtel, C. (2025). Students evaluating the trustworthiness of historical sources and Internet sources: A comparison. British Educational Research Journal, 51(2), 665–686. https://doi.org/10.1002/berj.4095

Warner, A., & Russo, A. (2026). Outdated notion? Teaching scholarly articles as the gold standard. College and Research Library News 87(1). https://crln.acrl.org/index.php/crlnews/article/view/27142/34994

---

About the Author

Sara Wink has been teaching undergraduates at PG since 2009. She loves to show her students that their words and voices have the power to make positive changes in their unique communities. One never stops learning how to write!