AI In Education Excerpt 2: What Constitutes Proper AI Usage and Plagiarism?
This article is the second in a series about AI and education by Andrew Rosston. Future excerpts will include discussion of AI in tutoring and safe AI usage. Read the first article, AI As a Learning Aid.
One of the most essential duties of teachers and parents in this early stage of AI adoption will be proper instruction of how to use such tools for students. Just as teachers in the 2000s warned not to immediately click on the first source on Google, fully trust Wikipedia, or cite personal websites as fact, they will need to show students how to find reliable information using AI and how to avoid plagiarism.
Students will need to learn not to directly cite or copy AI generations and instead how to delve more deeply, ensuring that the work is their own. As before, assignments involving heavier analysis will help students think critically and give answers in their own words. Anyone can copy a summary from a book in the library, but a student’s unique perspective on a topic cannot be idly reproduced.
Tools exist to detect both AI and traditional plagiarism. Strict honor codes and supervision may be necessary at times, with special consideration given to how examinations are administered to ensure students are tested on their own knowledge and skills. While some schools may completely forbid AI or require citations of every AI output used in research, others will allow AI to be used for brainstorming and research, as is already common in many colleges.
Teachers should ensure that, whatever their policies, students have a clear understanding of what constitutes plagiarism and how such tools are to be cited and referenced. For example, Oregon State University has formulated suggestions for AI policies but does not have a universal policy between classes and colleges; other institutions might take a more universal approach.
School-wide policies will need to be clear on how such requirements affect different subjects, as uses such as AI translation might have different impacts in history and foreign language classes. It may be more important to have a clear policy than what exact line is drawn between acceptable and unacceptable usage.
Teachers and professors may already be familiar with programs like Turnitin that use algorithms to detect similarities between student papers and other materials in order to prevent plagiarism. Essays can be fed into such programs to detect whether text was copied, or copied with small edits, from other sources. Programs like these can even flag source titles or organizations that commonly are written in other materials.
Now, with the genesis of AI large language models (LLMs), these programs face new sources of plagiarism and a potential new tool in combating it. For example, Turnitin now has several references to AI on its front page despite having operated since 1998, and it has shown a near-perfect detection rate for AI-generated writing, shared by several other programs.
It is good news for educators that plagiarism detection is rapidly becoming AI-focused, both in utilizing AI in this fight and in countering its abuses. Even Grammarly, a free writing assistance tool started in 2009, now sports an AI-powered plagiarism checkers as well as using generative AI to correct grammar and give feedback. (Anecdotally, the Grammarly program struggles with more advanced sentence structures, which are grammatically correct but likely less common in its training data. This may be typical, if AI training trends toward the “mean” of its dataset and prunes outliers.)
In extending their existing algorithms and machine learning work to utilize the more advanced versions of machine learning that power LLMs, such tools have a natural advantage in adapting to this new age of AI. For some educators, programs already in use may be sufficiently upgraded to detect ChatGPT generations, though it is important to verify which programs have such capabilities.
Andrew Rosston is a Business Analyst at OnlyMoso USA. He holds a B.A. in Business and Managerial Economics from Oregon State University.
