Teaching Fact-Checking Through Deliberate Errors: An Essential AI Literacy Skill

April 23, 2025

Abstract

This teaching resource explores an innovative pedagogical approach for developing AI literacy in a postplagiarism era. The document outlines a method of teaching fact-checking skills by having students critically evaluate AI-generated content containing deliberate errors. It provides practical guidance for educators on creating content with strategic inaccuracies, structuring verification activities, teaching source evaluation through a 5-step process, understanding AI error patterns, and implementing these exercises throughout courses. By engaging students in systematic verification processes, this approach helps develop metacognitive awareness, evaluative judgment, and appropriate skepticism when consuming AI-generated information. The resource emphasizes assessing students on their verification process rather than solely on error detection, preparing them to navigate an information landscape where distinguishing fact from fiction is increasingly challenging yet essential.

Here is a downloadable .pdf of this teaching activity:

Introduction

In a postplagiarism era, one of the most valuable skills we can teach students is how to critically evaluate AI-generated content. This can help them to cultivate meta-cognition and evaluative judgement, which have been identified as important skills for feedback and evaluation (e.g., Bearman and Luckin, 2020; Tai et al., 2018). Gen AI tools present information with confidence, regardless of accuracy. This characteristic creates an ideal opportunity to develop fact-checking competencies that serve students throughout their academic and professional lives.

Creating Content with Strategic Errors

Begin by generating content through an AI tool that contains factual inaccuracies. There are several approaches to ensure errors are present:

  • Ask the AI about obscure topics where it lacks sufficient training data
  • Request information about recent events beyond its knowledge cutoff
  • Pose questions about specialized fields with technical terminology
  • Combine legitimate questions with subtle misconceptions in your prompts

For example, ask a Large Language Model (LLM), such as ChatGPT (or any similar tool) to ‘Explain the impact of the Marshall-Weaver Theory on educational psychology’. There is no such theory, at least to the best of my knowledge. I have fabricated it for the purposes of illustration. The GenAI will likely fabricate details, citations, and research.

Structured Verification Activities

Provide students with the AI-generated content and clear verification objectives. Structure the fact-checking process as a systematic investigation.

First, have students highlight specific claims that require verification. This focuses their attention on identifying testable statements versus general information.

  • Next, assign verification responsibilities using different models:
  • Individual verification where each student investigates all claims
  • Jigsaw approach where students verify different sections then share findings
  • Team-based verification where groups compete to identify the most inaccuracies

Require students to document their verification methods for each claim. This documentation could include:

  • Sources consulted
  • Search terms used
  • Alternative perspectives considered
  • Confidence level in their verification conclusion

Requiring students to document how they verified each claim can help them develop meta-cognitive awareness about their own learning and experience how GenAI’s outputs should be treated with some skepticism and gives them specific strategies to verify content for themselves.

Teaching Source Evaluation: A 5-Step Process

The fact-checking process creates a natural opportunity to reinforce source evaluation skills.

As teachers, we can guide students to follow a 5-step plan to learn how to evaluate the reliability, truthfulness, and credibility of sources.

  • Step 1: Distinguish between primary and secondary sources. (A conversation about how terms such as ‘primary source’ and ‘secondary source’ can mean different things in different academic disciplines could also be useful here.)
  • Step 2: Recognize the difference between peer-reviewed research and opinion pieces. For opinion pieces, editorials, position papers, essays, it can be useful to talk about how these different genres are regarded in different academic subject areas. For example, in the humanities, an essay can be considered an elevated form of scholarship; however, in the social sciences, it may be considered less impressive than research that involves collecting empirical data from human research participants.
  • Step 3: Evaluate author credentials and institutional affiliations. Of course, we want to be careful about avoiding bias when doing this. Just because an author may have an affiliation with an ivy league university, for example, that does not automatically make them a credible source. Evaluating credentials can — and should — include conversations about avoiding and mitigating bias.
  • Step 4: Identify publication date and relevance. Understanding the historical, social, and political context in which a piece was written can be helpful.
  • Step 5: Consider potential biases in information sources. Besides bias about an author’s place of employment, consider what motivations they may have. This can include a personal or political agenda, or any other kind of motive. Understanding a writer’s biases can help us evaluate the credibility of what they write.

Connect these skills to your subject area by discussing authoritative sources specific to your field. What makes a source trustworthy in history differs from chemistry or literature.

Understanding Gen AI Error Patterns

One valuable aspect of this exercise goes beyond identifying individual errors to recognizing patterns in how AI systems fail. As educators, we can facilitate discussions about:

  • Pattern matching versus genuine understanding
  • How training data limitations affect AI outputs
  • The concept of AI ‘hallucination’ and why it occurs
  • Why AI presents speculative information as factual
  • How AI systems blend legitimate information with fabricated details

Connect these skills to your subject area by discussing authoritative sources specific to your field. What makes a source trustworthy in history differs from chemistry or literature.

Practical Implementation

Integrate these fact-checking exercises throughout your course rather than as a one-time activity. Start with simple verification tasks and progress to more complex scenarios. Connect fact-checking to course content by using AI-generated material related to current topics.

Assessment should focus on the verification process rather than simply identifying errors. Evaluate students on their systematic approach, source quality, and reasoning—not just error detection.

As AI-generated content becomes increasingly prevalent, fact-checking skills are an important academic literacy skill. By teaching students to approach information with appropriate skepticism and verification methods, we prepare them to navigate a postplagiarism landscape where distinguishing fact from fiction becomes both more difficult and more essential.

References

Eaton, S. E. (2023). Postplagiarism: Transdisciplinary ethics and integrity in the age of artificial intelligence and neurotechnology. International Journal for Educational Integrity, 19(1), 1-10. https://doi.org/10.1007/s40979-023-00144-1

Edwards, B. (2023, April 6). Why ChatGPT and Bing Chat are so good at making things up. Arts Technica. https://arstechnica.com/information-technology/2023/04/why-ai-chatbots-are-the-ultimate-bs-machines-and-how-people-hope-to-fix-them/

Tai, J., Ajjawi, R., Boud, D., Dawson, P., & Panadero, E. (2018). Developing evaluative judgement: enabling students to make decisions about the quality of work. Higher Education, 76(3), 467-481. https://doi.org/10.1007/s10734-017-0220-3

Disclaimer: This content is crossposted from: https://postplagiarism.com/2025/04/23/teaching-fact-checking-through-deliberate-errors-an-essential-ai-literacy-skill/

________________________

Share this post: Teaching Fact-Checking Through Deliberate Errors: An Essential AI Literacy Skill – https://drsaraheaton.com/2025/04/23/teaching-fact-checking-through-deliberate-errors-an-essential-ai-literacy-skill/

Sarah Elaine Eaton, PhD, is a Professor and Research Chair in the Werklund School of Education at the University of Calgary, Canada. Opinions are my own and do not represent those of my employer.


Celebrating 5 Years of Integrity Hour in Canadian Higher Education

March 31, 2025

Five years ago we started Integrity Hour, an online community of Practice by and for Canadian higher education #AcademicIntegrity enthusiasts, professionals, educators, researchers, and students. 

Today we had our five-year celebration, which also served as a closure of sorts. After serving as a co-steward of the community almost since the beginning, Dr. Beatriz Moya has started the next chapter of her career. 

We are working with some of our long-standing partners to reconceptualize what the next iteration of Integrity Hour will look like. For now, we will take a little pause as we regroup.

At our anniversary celebration meeting today, Brooklin Schneider encouraged us to share this guide widely, so we are posting it here, as an open access resource: “Integrity Hour: A Guide to  Developing and Facilitating an Online Community of Practice for Academic Integrity”.

Our collective outputs have been collaboratively conceptualized and co-developed. Here are a couple of other resources we have worked on over the years:

Reflections on the first year of Integrity Hour: An online community of practice for academic integrity

Academic Integrity Leadership and Community Building in Canadian Higher Education

In my remarks today I shared that being part of this weekly community of practice has influenced and informed my thinking, advocacy, and practice in way I could never have imagined. 

My gratitude to everyone who has been part of our community, sharing wisdom, knowledge, and resources. What an incredible half a decade it has been!

________________________

Share this post: Celebrating 5 Years of Integrity Hour in Canadian Higher Education – https://drsaraheaton.com/2025/03/31/celebrating-5-years-of-integrity-hour-in-canadian-higher-education/

This blog has had over 3.7 million views thanks to readers like you. If you enjoyed this post, please ‘Like’ it using the button below or share it on social media. Thanks!

Sarah Elaine Eaton, PhD, is a Professor and Research Chair in the Werklund School of Education at the University of Calgary, Canada. Opinions are my own and do not represent those of my employer. 


10 Recommendations for Academic Integrity Action: An AI-Assisted Experiment

January 27, 2025

I worked with a GenAI LLM (Claude) to create an infographic based on this open access article that I wrote:

Eaton, S. E. (2022). New priorities for academic integrity: equity, diversity, inclusion, decolonization and Indigenization. International Journal for Educational Integrity, 1-12. https://doi.org/10.1007/s40979-022-00105-0 

Here’s the result:

An AI-generated infographic with 10 circles, each of a different colour, with black text inside each of them.
AI-Generated infographic summarizing key ideas in a research article.

I wanted to see how #AI could be used to help with knowledge mobilization and communicating complex research ideas in simple and visually appealing ways. I am not 100% happy with this, but I am also not entirely unhappy. I am curious to see how researchers and scholars can use AI to help make our research more digestible to the general public.

At the end of the experiment, I asked Claude: Based on our interaction just now, what would you say are the benefits and drawbacks of working with AI to create infographics to share findings of research articles?

Claude’s reply:

“I can identify several key benefits and drawbacks of using AI to create infographics for research articles:

Benefits:

  1. Rapid Prototyping and Ideation
  • Quick generation of multiple concept ideas (as shown when I suggested 4 different infographic approaches)
  • Fast iteration based on feedback (like when we added your name and citation)
  • Ability to quickly visualize complex concepts from text
  1. Structured Organization
  • Systematic breakdown of complex research findings into visual elements
  • Logical arrangement of information hierarchies
  • Consistent application of design principles
  1. Customization and Flexibility
  • Ability to modify designs based on specific requests
  • Easy incorporation of branding elements or author attribution
  • Adaptable to different formats and styles
  1. Technical Precision
  • Exact color codes and measurements
  • Consistent spacing and alignment
  • Precise vector graphics that maintain quality at any size

Drawbacks:

  1. Format and Output Limitations
  • Cannot directly generate common file formats like JPG or PowerPoint
  • Unable to provide downloadable files
  • Limited to specific output formats (in our case, SVG)
  1. Technical Barriers
  • Users need to know how to convert or adapt the outputs
  • May require additional tools or software to get desired final format
  • Some technical knowledge needed to implement suggestions
  1. Creative Constraints
  • Limited to certain types of visual representations
  • May not capture nuanced design choices that a human designer would make
  • Restricted by pre-programmed design patterns
  1. Implementation Challenges
  • Gap between AI suggestions and practical implementation
  • Need for manual conversion or recreation in other software
  • Extra steps required to get to final usable format
  1. Quality Control Requirements
  • Need for human oversight and verification
  • May require professional designer input for polishing
  • Could need additional refinement for professional publication

For optimal results, it seems the best approach would be to use AI as a rapid prototyping and ideation tool, then work with a professional designer or use appropriate software to create the final, polished version in the desired format. This combines the efficiency of AI for initial concept development with the precision and format flexibility of professional design tools.”

Note: The numbering issue is a function of formatting in WordPress. I’ll fix that later.

This was an interesting experiment. I’ll be interested to see how AI apps can help us better communicate our research and scholarship to members of the general public.

________________________

Share this post: 10 Recommendations for Academic Integrity Action: An AI-Assisted Experiment – https://drsaraheaton.com/2025/01/27/10-recommendations-for-academic-integrity-action-an-ai-assisted-experiment/

This blog has had over 3.7 million views thanks to readers like you. If you enjoyed this post, please ‘Like’ it using the button below or share it on social media. Thanks!

Sarah Elaine Eaton, PhD, is a Professor and Research Chair in the Werklund School of Education at the University of Calgary, Canada. Opinions are my own and do not represent those of my employer.


In Memory of J. Tim Goddard: More Than a Supervisor

January 23, 2025
A photo of J. Tim Goddard sitting on the steps of his artist studio. His black dog is sitting next to him.

The academic world has lost a remarkable educator and mentor with the passing of J. Tim Goddard, former University of Calgary professor and administrator who then went on to serve as  Dean of Education at the University of Prince Edward Island. Tim’s impact on higher education extended far beyond his administrative roles – he was a transformative figure in the lives of countless graduate students, myself included.

As my PhD supervisor from 2005 to 2009, Tim demonstrated an extraordinary ability to balance academic rigour with compassionate mentorship. He took a chance on my research into revenue-generating programs in higher education, recognizing the importance of examining how such initiatives could be administered ethically in an era of government funding cuts.

Tim’s dedication to his students was unwavering, even in the face of profound personal tragedy. In 2006, his daughter Captain Nicola Goddard was the first Canadian female soldier to be killed in combat during her time in Afghanistan. Despite this devastating loss, Tim continued to guide and support his graduate students with remarkable strength and grace.

What set Tim apart was his practical wisdom, delivered through memorable maxims that still guide many of us today. “This isn’t your magnum opus!” and “A good thesis is a finished thesis!” were frequent reminders to stay focused on completion rather than perfection. He had an uncanny ability to guide students back to their research path whenever they strayed too far into tangential explorations.

His mentorship extended beyond the academic realm. When I traveled to Charlottetown to write my thesis, Tim and his wife Sally opened their home to me – a gesture typical of how Tim viewed his students as extended family. Yet this warmth never compromised his high standards; his insistence on impeccable writing, citations, and APA formatting shaped many of us into better scholars.

J. Tim Goddard and Sarah Elaine Eaton in PEI. There is a lighthouse in the left background.

After retiring in 2018, Tim embraced new creative pursuits in painting and novel writing, approaching these endeavours with the same passion he brought to academia. His legacy lives on through the countless educators and researchers he mentored, who continue to apply his teachings in their own work and supervision of the next generation.

Tim Goddard understood that a PhD journey was not just about producing research, but about preparing for a career. He showed us how to be not just scholars, but mentors, teachers, and advocates for our own students. His influence will continue to ripple through generations of educators, researchers, and leaders in higher education.

________________________

Share this post: In Memory of J. Tim Goddard: More Than a Supervisor – https://drsaraheaton.com/2025/01/23/in-memory-of-j-tim-goddard-more-than-a-supervisor/

This blog has had over 3.7 million views thanks to readers like you. If you enjoyed this post, please ‘Like’ it using the button below or share it on social media. Thanks!

Sarah Elaine Eaton, PhD, is a Professor and Research Chair in the Werklund School of Education at the University of Calgary, Canada. Opinions are my own and do not represent those of my employer. 


The GenAI Gender Gap

January 10, 2025

There is a gender gap when it comes to GenAI.

Just 26.3% of the European Union’s artificial intelligence (AI) professionals are women, according to a report from LinkedIn.

In my work with of the Women for Ethical AI (W4EAI) UNESCO platform, we had similar findings in our gender outlook study.

An AI-generated image of a group of women.

There are no easy solutions to this gap, but for those working in this area, some five concrete things you can do to promote gender inclusion (and equity in general) are:

  • 
Invite women into leadership roles, strategic planing for artificial intelligence and advanced technology.
  • Ensure that policies explicitly include women, girls, and other equity-deserving groups.
  • Invite women (and in particular, early career women and those who are precariously employed) to share and showcase their expertise and knowledge (and compensate them for their contributions).
  • Create formal sponsorship programs for women and girls who want to develop their knowledge and cp-competencies related to AI, with ongoing opportunities for learning and skill development.
An AI-generated image of a group of women.

There are a myriad of ethical complexities when it comes to artificial intelligence and gender is only one of them. Acknowledging inequalities and then working to support equity, fairness, and justice will remain ongoing work in the years to come.

References

AI in the EU: 2024 Trends and Insights from LinkedIn. (2024). https://economicgraph.linkedin.com/content/dam/me/economicgraph/en-us/PDF/AI-in-the-EU-Report.pdf

United Nations Educational Scientific and Cultural Organization (UNESCO). (2024). UNESCO Women for Ethical AI: Outlook study on artificial intelligence and gender. https://unesdoc.unesco.org/ark:/48223/pf0000391719

________________________

Share this post: The GenAI Gender Gap – https://drsaraheaton.com/2025/01/10/the-genai-gender-gap/

This blog has had over 3.7 million views thanks to readers like you. If you enjoyed this post, please ‘Like’ it using the button below or share it on social media. Thanks!

Sarah Elaine Eaton, PhD, is a Professor and Research Chair in the Werklund School of Education at the University of Calgary, Canada. Opinions are my own and do not represent those of my employer.