Speaker 1: Hello, I'm Dr. Norma Jones. In this video, I will offer some quick tips to help teachers uphold academic integrity and plagiarism as artificial intelligence, or AI, continues to proliferate and impact our educational sphere. To address these issues, we will discuss the following quick tips. How traditional plagiarism checkers are not as effective in finding AI-generated text. Then we will discuss holistic and creative ways of addressing academic integrity and AI plagiarism by first considering our policies. We will move on to how we can demonstrate and model appropriate AI usage. Additionally, we'll look at recreating assignments that are less susceptible to AI proliferation issues. We will then discuss approaches to students who appear to be violating AI usage policies. Let's start with plagiarism checkers. In the recent past, plagiarism checkers were effective because they compared students' submitted text against a massive database of papers, websites, and other sources to find direct matches and paraphrased content. But here's the catch. These plagiarism checkers rely on these databases. If the AI-generated content is not on a database, then it is harder, if not impossible, to detect. For instance, AI models such as ChatGPT create unique content that isn't copied from any source, making it difficult for plagiarism checkers to identify. Plus, AI is adaptable. It can create unique responses based on specific prompts and contexts. This means each piece of AI-generated text can be unique, which may allow that text to slip through the cracks of traditional plagiarism checkers. From there, some new tools were created to find AI-generated content. These AI detectors often scan for two elements. The first is perplexity, and the second is burstiness. Perplexity is a way to gauge how well a language model predicts a piece of text. Think of it as measuring how surprised the model is by the words it's trying to predict. For example, the sun rises in the east would score low on the perplexity scale, whereas the sun eats a sandwich would score higher. Let's move on to the second parameter. Burstiness is all about those rapid, intense bursts of activity that happen in a short period. When we're talking about AI-generated text, burstiness describes how often certain words or phrases pop up, and how they're spread throughout the text, along with variations in sentence length and structure. However, as AI learns more and more from us, AI can write and rewrite text that sounds just like a human wrote it, following our styles and patterns. For example, OpenAI stopped developing their own AI detector within months because of the number of false positives. As a result, it means that as educators, we need to understand the limitations of traditional plagiarism checkers, AI detectors, and keep up with the evolving AI tools. Next, we also must include holistic and creative ways to help us better serve students as AI continues to proliferate academe. Let's start with... Our policies. First, we need to review our current academic integrity policies. It's crucial to make sure they're up to date with the latest tech trends, especially AI. Next, we must develop guidelines on AI usage in assignments. We need to spell out what's acceptable and what's not. It will be helpful to be clear and specific. Currently, across institutions, we have a few different levels of AI usage policies. Some districts have zero usage policies, in that students are not allowed to use any AI for any coursework. If you choose this type of zero usage policy, consider how you will address writing assistance, such as grammarly and auto-AI generated content across browsers such as Google. Also, this stance may hinder student success outside of academia, as companies are hiring based on applicants' abilities and competencies in using AI. Other institutions allow some usage, but that could be based on a percentage, specifics, or tasks about how AI should be used. For example, in my classes, students may use AI to brainstorm ideas, help with understanding materials, and polish drafts. Regardless of the amount of AI we allow or disallow, we must clearly communicate these policies to our students. They need to understand the expectations and the consequences of violating these academic integrity rules. Next, let's consider our role as teachers to model and demonstrate appropriate AI usage. It is essential for us as educators to demonstrate proper AI usage in our own work. By doing this, we set a positive example for our students. We can incorporate demonstrations in class on how AI can be a powerful tool to aid learning and research, rather than a shortcut to completing assignments unethically. We also need to discuss the ethical considerations of using AI. It's important for students to understand the value of maintaining integrity in their academic work. For example, I tell students that if they replace their learning with AI now, AI will not be taking their jobs in the future. Instead, the students themselves will be giving away their jobs to AI. Next, let's look at how we can recreate assignments that are less vulnerable to AI misuse. First, we should focus on creating more authentic assessments. By designing assignments that require critical thinking, individual experiences, and unique insights, these elements are difficult for AI to replicate. Next, break assignments down into steps, as scaffolds. These scaffolded assignments allow students to submit parts of their work, in stages. This process helps students learn, and makes it harder for students to let AI take over the entire process. We can also look at the students' work in stages, to compare their original work with any potential AI misusage. Also, explore creative assignment formats, like multimedia projects, oral presentations, and collaborative work that AI tools can't easily handle. Lastly, use self-reflections and evaluations as part of assignments. This personal touch not only fosters deeper learning, but also makes it harder for AI to generate authentic, individualized content. Let's move on to approaching students that may have violated your AI usage policies. You can start with an initial conversation with the student to discuss your concerns, and give them a chance to explain their work. For example, I tell students that their submission has indicators of violating our AI usage policies. Then, I ask them to meet with me to discuss their submission. Next, make sure you gather evidence. Do not assume. For example, if you scaffold assignments, you could compare the suspected work with the student's previous submissions. We must be careful here because we have seen cases where false positives have led to multiple student complaints. Also, multiple institutions have recommended that instructors not use AI detectors. Next, as AI is still a newer technology for us, consider educational interventions like assigning workshops on academic integrity, or redoing the assignment instead of using punitive measures. Lastly, it is crucial to enforce these policies consistently to maintain fairness and uphold our academic standards. Before I conclude this video, let me address our roles as educators in helping students better learn how to use AI appropriately. I demonstrate and use AI in classes. I also hold regular workshops on my campus for students, faculty, and the broader college community. As a result, my students reported positive outcomes. First, my students feel that my usage is appropriate and model their usage on mine. This way, they start with a positive and ethical example. Second, they reported that they do not consider using AI to cheat, because they have the tools to use AI correctly. In other words, they learn the appropriate, transparent, and ethical guidelines that I model, and they follow suit. Finally, outside of academe, students are reporting that they are getting jobs and promotions because they have learned how to use AI appropriately and ethically. Let's review. We discussed upholding academic integrity by addressing AI in these areas. We discussed how traditional plagiarism checkers are not as effective in finding AI-generated text. We discussed holistic and creative ways of addressing AI proliferation. We first looked at AI usage policies. Then, we addressed demonstrating and modeling appropriate AI usage. Next, we looked at recreating assignments that are less susceptible to AI proliferation issues. We then reviewed approaches to students who appear to be violating AI usage policies. We concluded with some benefits. I'm Dr. Norma Jones. Thank you for watching. Stay tuned and subscribe to my channel for more quick tips on AI usage in academia.