Informed Pedagogy is the Key to Solving #AI Plagiarism

Recently, we’ve been hearing a lot about ChatGPT by OpenAI and, more broadly, the potential impact of artificial intelligence (AI) on higher education.  

eCampusOntario’s Research and Foresight team has been closely following this topic for the last year exploring how #AI and similar technologies can increase human skillset and potential. For example, how these technologies create new employment opportunities, improve access to on demand self-paced training, and create spaces to practice and develop foundational employment skills. 

However, as with any technological advancement, AI presents opportunities and challenges for postsecondary education. A new issue sparking attention? The increase in plagiarism from learners submitting AI generated work in place of their own.   

Authenticity and Assessment – What’s Changed?

Global News recently examined how ChatGPT has changed education in the short time that it has been on the scene and has educators rethinking their approach to assessment. A couple of highlights include:  

• Human beings could only identify text authored by AI 52% of the time, which is only slightly better than guessing. 
• Perceptually, some learners may not view the use of ChatGPT as cheating, but rather as another common tool like Google search. 
• Valued at $29 billion USD, ChatGPT will soon benefit from a $10 billion US investment by Microsoft, suggesting that the AI will only improve and get harder to detect. 

ChatGPT’s capabilities pose a challenge to both educators and plagiarism detection software. But what are the safeguards for this emerging phenomenon? The technology world is already tackling the issue with improved tools that can identify AI authored content.   

• Recently launched GPZero is supporting academic honesty by leveraging AI tech to identify plagiarized assignments.   
• Turnitin indicates that their platform can currently detect AI generated writing (A form of writing which tends to have numerous factual errors, uses linear language models, and broad word choices that produce signals detectable by anti-plagiarism tools)  
• OpenAI acknowledges their responsibility to ensure honorable use of their product and are considering the addition of cryptographic signals to their AI content that is easily detectable by software like Turnitin. 

Notwithstanding the above, detection tools have a long way to go, with many solutions failing to identify AI authored content. 

How Can Educators Adjust? 

By its very nature, the capability of AI to write more human-like text will continue to improve as it learns from continuous use, making detection increasingly difficult. In response, many higher education institutions are rethinking their academic integrity policies and approach to assessment. 
 
So, what can educators do? Below is a comprehensive list of some strategies that leverage authentic assessment and provide learners with the skills to avoid academic dishonesty. 

• Talk to your learners about the importance of original work as part of their own learning and development.  It is important that students understand that learning does not occur from completion of a final product, but rather is a holistic process of working through content and learning along the way. The term ‘Plagiarism’ should be clearly defined in a written document. Educators should consider developing co-created honour codes with learners so learners know how to avoid plagiarising. 
• Redefine academic integrity policies to explicitly address artificial intelligence and its inappropriate use in academia. Importantly, educators should clearly indicate the purposes for which AI can be legitimately used. 
• Create milestones and checkpoints in major written assignments at key points in the writing process: Annotated Bibliography, thesis statement, reference list, outline, research notes, and final essay. Scaffolding the creation of assignments helps learners plan and chunk work while making the use of AI enhancements unnecessary.   
• Encourage accountability through peer editing and group assignments. 
• Use writing prompts that require analysis and critical thinking rather than facts or require that responses connect with previous class discussions or readings.  These prompts focus more on the application of knowledge in real-world contexts as opposed to memorization of content and connects to resources unavailable to any AI. 

• Augment essay and report assignments with presentations.  This will require learners to explain their work and answer questions in real-time. 
• Consider alternative assessments to traditional essays. Suggestions include case study analysis, creation of infographics, in-class timed written assessments, podcasts, game creation, and more. Explore Open Education Resources to discover how educators are engaging learners with different kinds of assessments. A few OER suggestions include Write it Again OER, Transforming Assessment: Strategies for Higher Education, and Beyond the exam: An alternative online assessment toolkit 

Preparing Learners For the Future

The University of Toronto acknowledges the concerns around AI and academic dishonesty while sharing an aura of optimism; AI presents an opportunity for educators to prepare learners for the evolving digital age of information. Modeling effective and honest use of AI will help prepare learners for the world beyond school.  
 
This will require postsecondary educators to mitigate the misuse of AI in academia, correct mis-information about the technologies, and find ways to leverage these tools as teaching and learning solutions. Educators at University of Windsor and Western University are carefully thinking about what AI means to teaching and learning while acknowledging that there is still lots to learn. They note that it is important to consider how to use AI tools appropriately and honestly.  

How can we integrate AI into our programs? Deliberately considering your assessment options and ways to use AI can help you creatively navigate the impact of AI on your practice. Some suggestions for using AI in teaching include generating AI content and having learners peer edit, source, or critique the work. Additionally, AI can emulate conversations to help learners practice communication skills or improve engagement in online learning.  

Your Turn: How Have You Dealt with AI in Your Practice?  

Join the conversation and comment below. We would be happy to hear your experience with how AI is impacting teaching and learning in your institution, as well as listening to your strategy to meet this new challenge.  

Bio: 

Don Eldridge is a Digital Learning Associate with eCampusOntario where he investigates the use of adaptive learning and other innovative technologies in higher education.  Visit our Adaptive Learning webpage to discover webinars and more information.