Generative Artificial Intelligence

Generative AI offers numerous benefits that can greatly enhance the learning experience for our students. Here are just a few examples of the potential of these technologies.

Individualized Student Support: AI-powered tools can provide instant feedback for students on their own assignments, highlighting areas for improvement and offering explanations for correct answers. Generative AI can also identify areas of struggle and provide targeted resources or exercises to address specific challenges, ensuring a more effective use of student’s study time. The immediate feedback loop provided by AI accelerates the learning process and encourages iterative learning.

AI-driven chatbots and virtual assistants can offer round-the-clock support, answering students' questions and providing assistance outside of traditional classroom hours. This ease of access promotes continuous learning and self-directed study. AI algorithms can also suggest relevant learning resources, such as articles, videos, and textbooks, based on a student's preferences and learning objectives, promoting self-directed learning.

Enhancing Creativity: AI-generated prompts can inspire students' creative thinking and encourage them to explore new ideas. By collaborating with AI in generating ideas and even initial drafting, students can push the boundaries of their imagination and innovation.

Inclusive Learning: AI tools can provide accessibility features, such as text-to-speech and translation services, enabling students with diverse abilities and backgrounds to access educational content more effectively. AI-powered language learning platforms can provide real-time language correction, pronunciation feedback, and vocabulary suggestions, making language acquisition more interactive and effective for ESL learners.

GenAI literacy, or the ability to understand and navigate the world of generative artificial intelligence, will become increasingly important in today's technologically-driven landscape. Just as traditional literacy empowers individuals to communicate effectively and critically evaluate information, GenAI literacy empowers students to harness the capabilities of AI tools responsibly and ethically. With the growing integration of AI in various aspects of our lives, from content creation to decision-making, GenAI literacy ensures that individuals can discern AI-generated content, make informed choices about its use, and maintain human creativity and authenticity. As AI and automation reshape industries, exposure to generative AI tools equips students with valuable digital literacy skills and prepares them for a technology-driven job market.

We are in the early stages of harnessing artificial intelligence to enhance education. As these technologies continue to evolve, we’ll see the emergence of more capabilities, greater accuracy and a growing reliance on these tools in support of teaching and learning.

Content Creation: GenAI can aid instructors in the development of high-quality course materials. It can generate summaries, outlines, and study guides, condensing complex topics into digestible formats. Additionally, GenAI can assist in creating visually appealing presentations and multimedia elements that enhance the learning experience.

Automated Assessment: AI can automate the grading process for assignments and exams, freeing up instructors' time for more meaningful interactions with students. When disclosed and appropriate for the course goals, this allows educators to focus on providing targeted feedback and support.

Enhanced Engagement: AI tools, such as chatbots and virtual assistants, can engage students in real-time discussions, answer questions, and provide support outside of classroom hours, enhancing engagement and accessibility.

Data-Driven Insights: AI can analyze vast amounts of data to identify trends and patterns in student performance. Educators can use these insights to make informed decisions about curriculum design, teaching strategies, and interventions. By analyzing historical data, AI can predict students' potential areas of struggle and offer early interventions to prevent learning gaps from forming.

Real-World Applications: AI can simulate real-world scenarios, enabling students to practice problem-solving and decision-making in a safe virtual environment, particularly in fields like healthcare, engineering, and business.

Career Development: GenAI can assist both students and educators in career development. It can help students create well-structured resumes, cover letters, and portfolios that showcase their skills and experiences. GenAI can provide insights on industry trends, job market demands, and required skills, aiding students in making informed career decisions. Educators can also use GenAI to stay updated on emerging trends and tailor their curricula to meet evolving industry demands.

While AI holds great potential to revolutionize higher education, its implementation should be guided by ethical considerations and a focus on preserving the human aspects of teaching and learning. By effectively integrating AI into educational practices, we can create more engaging, personalized, and effective learning experiences for our students.

Before making any determinations about how AI could support your teaching and learning it is advisable to spend some time familiarizing yourself with a few of the common tools and the general capabilities of generative artificial intelligence. We encourage you to consider how these emerging tools align with your overall teaching philosophy and the learning objectives of the courses you teach. We endorse a student-centric approach to the adoption and use of these tools within your coursework. GenAI will have tremendous impact across most disciplines so it’s in our student’s best interest to prepare them for a future where they can use these tools effectively and ethically.

Now is a great time to carefully consider how you are assessing student learning in your courses. If your stated learning objectives do not include improving the student’s ability to produce written work, then you have some flexibility in how you approach the use of GenAI tools. You can alter assignments to make them less friendly to the use of AI text generators like ChatGPT. See the section below on Using Alternative Assignments and Assessments. If you are attempting to grade students on their writing abilities, you have a greater challenge. We offer guidance below in the section titled Rethinking Current Writing-Based Assignments.

As you think about how to incorporate the potential of GenAI into your coursework you should keep in mind the following considerations.

Student Rights: Instructors do not have complete freedom in their course policies because they are beholden to the student code and the FAIR system. It is possible that you could unintentionally violate student rights, particularly if you do not follow the due processes afforded to students through the student code and the FAIR system. To process plagiarism as an academic integrity issue, you must submit proof in support of the allegation, and it might not always be possible to have proof in writing of AI use. AI detection programs have already been shown to be biased and inaccurate and are unlikely to be persuasive in an academic integrity case. Additionally, students have a right to use these tools to enhance their own learning. By being unnecessarily punitive in your policies you risk negatively impacting your student’s academic growth and understanding of these tools.

Equal Access: If you are creating assignments that endorse or require the use of GenAI capabilities, be sure that all your students have easy access to these tools. Many of the options are free, but students with greater access and the ability to pay for more advanced and updated tools may have unintended advantages not available to all your students. Group projects could be used to help minimize this concern.

Accuracy and Bias: It is important to educate your students about the potential for inaccuracy and the inherent bias within these tools. AI algorithms can inadvertently perpetuate biases present in the training data potentially leading to discriminatory results. These tools can also generate incorrect, and at times seemingly random, data that is presented as fact, including sources and citations that can appear factual. It is important we assist our students in becoming savvy AI users to avoid replicating these biases and spreading inaccurate information.

Difficulty in Policing AI Use: While it may be tempting to try to set strict policies against all AI use in your course, this puts an undue burden on you and your teaching assistants. It is difficult to positively detect the use of AI generated text within a written assignment. The available tools are simply not effective in providing the evidence needed to build an academic integrity case against a student (see the Limitations of AI Detection Tools below). Our pedagogies should be built with critical AI literacy in mind, so it’s important to think through what goals AI prohibition is going to meet and whether enforcement is how you want to spend your time and energy.

Teach your students about academic integrity:

Academic Integrity is a critical issue, and we should address the concept with our students and include it in the syllabus. Your syllabus policies should address the importance of academic integrity and the use of AI. Avoid punitive language which can break trust. Stress the positive.

Students should be made aware of the benefits of academic integrity along with the consequences. It is often better for the student and instructor to focus on ways to improve learning and disincentive cheating overall. Explain to your students the intention of your policies and focus on the outcome you are seeking. Highlight how your course policies are intended to help them succeed and how using AI would defeat the intended goals of the course.

It is also important to consider why students might “outsource” their work and address those issues. Often, time management skills are to blame, or the student does not know how to get started with a project. You can assign students to speak to a librarian about a topic, provide them with instructional resources, or structure assignments in such a way that they complete bite-sized, manageable components over time, and do not get in a rush at the end.

Share your Experience:

As a campus, we are all exploring the future of artificial intelligence and its impact on higher education. We encourage you to share your experiences, both good and bad, with us here at CITL, with your colleagues and with your students. See the Illinois Community section below for more information about how you can share your insights and learn from others on our campus.

The Gies College of Business produced an infographic of Generative AI Guiding Principles that streamlines the considerations we’ve discussed above. This is a helpful resource for educators interested in using AI effectively and ethically in their teaching, while becoming more aware of possibilities and challenges of this powerful emerging technology.

Using artificial intelligence in a higher education environment brings numerous benefits, but it also raises important ethical considerations that should be addressed to ensure equity, transparency, and responsible use. Some key ethical considerations include:

Inherent Bias: AI algorithms can inadvertently perpetuate biases present in the training data potentially leading to discriminatory results. Remember that GenAI tools produce results that are representative of the original creators of the data used for training and can replicate any inherent biases in that population of creators. 

Academic Integrity: The use of artificial intelligence for assignments and assessments can blur the lines between original work and automated assistance. Faculty should clearly define the boundaries for using AI in student work to maintain academic integrity and discourage plagiarism. Students should be educated about the ethical use of AI tools and reminded that text and ideas generated by AI should be properly cited to avoid plagiarism. 

Privacy and Data Security: AI systems generally require students to register to gain access to the tool. Faculty should never require students to register for any platform using their official university email address. It is the faculty members’ responsibility to safeguard student data following all relevant regulations covered by FERPA (Family Educational Rights and Privacy Act). Students and faculty should also be aware that sharing sensitive or proprietary data with third-party AI tools may expose that data to potential security risks and unauthorized access.

Accessibility and Equity: Institutions should ensure that AI systems are compliant with accessibility standards and do not inadvertently disadvantage certain groups of students. 

Another important factor to consider is who has easy access to these technologies. The mandatory use of GenAI for assignments could worsen existing issues often known referred to as the digital divide.

Addressing these ethical considerations requires collaboration among educators, administrators, and students. By adopting a thoughtful and proactive approach, we can harness the benefits of AI while upholding ethical standards and ensuring a positive and inclusive learning environment.

Production and Spread of False Information: Generative AI can sometimes produce outcomes termed "hallucinations." This is a phenomenon that occurs when AI generates incorrect, and at times seemingly random, data that is presented as "fact." Without human confirmation of AI-sourced information as accurate, a risk of the introduction and spread of misinformation occurs.

Cognitive Offloading: Some studies indicate that offloading cognitive process may impair long-term memory formation - consider the utilization of web mapping services for directions and search engines for recalling knowledge. Research into the impact of gen AI on cognitive processes and memory retention is a fairly new area of inquiry. Keeping updated on findings and developments may aid in determining if some knowledge acquisition should be completed with minimal or no cognitive offloading.

Re-think current writing-based assignments:

Traditional writing assignments pose the greatest challenge to faculty since the advent of ChatGPT, but this can be an opportunity to reconsider the core purposes of student writing, and to refocus your assignments in ways that make it harder for students to rely on AI. These tips may help:

• Offer highly specific prompts: Giving students prompts that relate to in-class discussions, online class forums, guest speakers, peer presentations, in-class debates, or something local will deter the effectiveness of Chat GPT.
• Make it personal: Make assignments more personalized and reflective by asking students to share the ways their lives intersect with the topic. Making it personal engages students more meaningfully in their writing and makes AI less useful.
• Add in-class writing sessions: Asking students to write for short periods during class can help measure comprehension and subject matter knowledge, although this is less useful for assessing various forms of writing, or for creative writing.
• Compare and contrast: Compare AI-generated and human text passages line-by-line, examining thesis statements, organization, evidence and support, arguments and logic, overall impact, and persuasiveness. Closer reading can help students see the limitations of AI.
• Ask for refinements in the written piece: Ask students to focus on rhetorical situations by adjusting for audience, purpose, voice, and tone. Students may see that it’s harder to get AI to write appropriately than it is to write specific assignments on their own.
• Use the tool for feedback: Ask students to input portions of their writing into ChatGPT to receive feedback on grammatical errors. AI can edit writing and explain what it changed and why which can help students learn.
• Focus on the writing process: Getting students to focus on the continual process of improvement helps them to listen for and refine their own voice. Focusing on the process more than the product helps them understand the rewards and challenges of becoming good writers.
• Scaffold assignments: Break big assignments into smaller parts—prewriting, drafting, revision-- and provide incremental feedback with lower-stakes grading. This helps students develop their skills more intentionally and you will also come to recognize their voices and styles.
• Teach proper citation methods: Help students learn how to cite the use of AI properly as they would any other source. Recommended style sources are MLA and APA. Discuss formatting issues and how to cite for different uses such as ideation, inspiration, or brainstorming.

Use alternative assignments/assessments:

Writing is not the only way to assess students. Alternative means of assessment provide students with rich and meaningful ways to demonstrate their learning. By being creative with a rubric, we can assess a variety of assignments fairly and equally. Giving the students options lets them be more engaged in their assignments.

While it may be impossible to make all your assignments and assessments AI-proof, some modifications can make using AI less effective for students. Assignments that are more hands-on and personal help ensure that students are describing and acting upon their own experiences. In class and online activities where students are asked to make things or participate with others helps connect course concepts to real-world experiences. Multi-modal assessments that require students to show what they know offer opportunities for more individualized and creative work. Modifying your course in the age of AI will take time and effort, but moving toward more engaged forms of learning and assessment can offer important benefits to students.

• Focus on in-class performance: Give short, low-stakes quizzes at the start of class to encourage students to do the reading. Alternate different types of quizzes with short in-class writing periods. Include surprise questions that ask them to reflect more deeply on a topic. Offering in-class participation points for these activities will make them seem less punitive while emphasizing the need to prepare for class, and keeping students focused on key concepts and ideas.
• Focus on individual presentations: Require students to make short informational presentations in class, alone or in small groups, tying related concepts together or demonstrating a problem-solving strategy. Even if they have harvested ideas from ChatGPT they will still have to explain and discuss them. Try rotating students through different roles as Leader, Fact Checker, Disrupter and Arbiter to provide alternative learning experiences, and use a question guide or rubric to help students formulate their ideas and anticipate questions from peers.
• Focus on applied learning in small groups: AI may be less useful when students are being asked to apply information in different formats. Having students engage in classroom roleplay exercises, debates, and panel discussions will force them to prepare ahead of time and reflect upon their understanding. Participating in Think-Pair-Share exercises, Fishbowl discussions, and Brainstorming sessions help students to process what they are learning while exploring different viewpoints.
• Focus on making learning visible: Ask students to show what they know by working alone and in groups to create detailed concept maps, diagram complex processes, complete timelines, build fishbone maps, Lotus diagrams, or hexagonal graphs to pull apart concepts and examine them in parts. Increase collaboration by having students answer questions or share ideas by posting sticky notes in the room or online. Explore digital debate and annotation tools to continue learning outside of class and retain an artifact that can be used later for review.
• Focus on making things: Ask students to create short podcasts or videos of themselves or their peers describing a concept, being interviewed, unboxing a product, or documenting the steps for completing a project or building a model. Have students work alone or in groups to create a game that teaches a concept, tells a story, or demonstrates what they have learned about a topic. Send them to makerspaces on campus to build a prototype, create an augmented reality poster, or explore other creative tools.
• Focus on using the tool more effectively: AI can offer students useful feedback and tutoring but we can take this a step further by asking them to document precisely how they use it, including how different programs work, what prompts they used, what results they got, and what they found that was missing or wrong in those results. Using AI to generate an initial outline for a paper or project is one thing but asking students to describe what they learned from this, and how they will develop it further is a challenging and worthwhile next step.

Instructors set policy for their own courses through their syllabi. CITL has some general guidance about creating a syllabus. However, it is also recommended that instructors directly address the use of generative AI in their course with a simple reminder that the use of artificial intelligence agents without the proper citation is a form of plagiarism.

Rather than a blanket AI policy in your syllabus, you may want to consider how different AI policies might be applicable for different assignments and assessments within your course. This strategy provides some flexibility in the utilization and support of AI as a learning tool, while still allowing you to impose strict limitations on its use when necessary.

Several colleges and departments have shared their guidance, below.

• Gies College of Business provides the following syllabus guidance.
• The College of Law provides sample syllabus statements for various levels of AI use.
• For Fall 2023, Rhetoric 233 uses the linked statement on Predictive Writing Technologies in their syllabus.
• The Center for Writing Studies has issued this draft statement on General Education Advanced Composition Courses and Generative Artificial Intelligence Tools for Fall 2023.
• Numerous instructors from a variety of educational institutions have crowdsourced their course policies on using generative AI and shared them for the greater good of the community. 

Do not depend on AI detection tools to address the use of generative AI in your assignments. They are unreliable and easy to circumvent with small amounts of editing. They also generate many false positives. Consider using these results as a conversation starter with students. Falsely accusing students of cheating with generative AI can quickly damage a learning community.

ASU & UCLA have both opted out of TII’s AI detection for now. Others like the University of Kansas caution faculty. The UM System has summarized their concerns and includes videos from Turnitin on how the detection works. Here’s a Washington Post article on the efficacy of Turnitin's detection and how the company views the feature to prompt a conversation with students instead of an accusation