Assignment Adjustments in Response to AI Text Tools
First, be sure that you:
- Get familiar. Right now, a version of ChatGPT is free to use, and other programs have free trials. Bing Chat is free with an email account and its Creative Mode (also free) is built on the paid version of ChatGPT. Take advantage and experiment. It is easier to spot AI-generated content after working with one of these machines for a day. Such writing machines will continue to improve and be less obvious to humans, so faculty members will have to rely on their own abilities. You should try to run your assignment through ChatGPT, review that answer, and tell your students about your experience (and that you’ve saved the output). Note: ChatGPT does not produce the same answer each time its posed.
- Make (and articulate) your policy. Since AI writing programs are relatively new, students may not know if using them is acceptable. A good policy is simple and concise for students, and it also gives you flexibility if you decide to work it into your curriculum. Simply making your students aware that you’re familiar with the technology decreases their likelihood of using it to cheat.
- Consider time and purpose. ChatGPT offers a huge benefit to those who are short on time or who perceive an assignment as low-value (this opinion can often be valid, based on their lives outside school or when evaluating their short or long-term priorities). Examine your course schedule to determine if you are unknowingly pushing students to take short-cuts.
- Review your course design. Consider whether the course unintentionally encourages the use of AI tools due to an overemphasis on grades or high-stress, high-stakes assignments. If a perfect product (test, paper) is the only way to receive an A, students are more likely to consider cheating. Instead, look to provide incentives for the behaviors and habits that are associated with strong learning—for trying—as well as producing. In your effort to fully assess student learning, make sure you assess (and reward) the processes that are needed to be a strong learner in your course: reading, speaking, improving, reflecting on one’s learning, etc. Review your class structure, grading criteria, and rubrics to make sure you’re setting your students up to adopt strong learning strategies.
Then, create assignments that:
- Increase the specificity of your question or task: For example, rather than an open-ended writing prompt (“compare X and Y”), ask for specific information or examples (ask students to expand on one point discussed in class, create a scenario and ask for specific resources/sources, etc.)
- Require course-based research. Interviews, surveys, experiments and observations are challenging for students, but they are difficult for AI to perform. However, these machines are capable of writing up fake results. To counter this, have students turn in any raw data or documentation as proof of their efforts.
- Require the application of key concepts to specific local cases, problems, or contexts: For example, provide students with a reading and ask them to analyze and expand on the topics and how they affect Union in order to focus the students’ thoughts and responses on that paper, example, or case study.
- Require inclusion, citation, and direct response to specific course texts: For example, prompt students to cite the textbook or another reading in class, in addition to X number of additional sources that they find in Y database. ChatGPT and its competitors were largely trained on open-access internet content. By using paywalled materials, these programs will struggle to produce meaningful responses from that content.
- Include multi-modal elements. ChatGPT is (currently) only capable of outputting text. And, while other AI applications can output video, audio, and images, putting them together is still difficult and time-consuming enough to probably be easier to just do manually. Consider adding presentations, audio or video interviews, and other modes of communication to your projects.
- Require drafts and revisions. Multiple, sequential stages of a larger assignment can be useful to help students scaffold their understanding as well as help them produce their own content. Peer reviews and iterative development, particularly when combined with a short (one-paragraph) reflection on what changes were made as a result of a first submission or peer review help students continue to frame and develop their learning.
- Promote In-Person Discussions and Activities. In-class activities that stimulate student interaction and deep engagement with the material are impossible for AI to mimic. Consider asking students to discuss a particular topic or participate in problem-solving sessions in class. Remember, the nuances of face-to-face interactions are yet to be replicated by AI.
- Nurture Original Thought and Deep Analysis. One way to circumvent the use of AI tools is by developing assignments that call for unique perspectives and critical thinking. Tasks that require personal reflection, interpretation, or innovative solutions to complex problems will motivate students to create original work rather than resorting to AI tools.
- Implement Peer Review Sessions. Reinforce the value of the peer review process in your course structure. Not only does this method support the growth of ideas and constructive critique, but it also engages students with their classmates’ work, making them less inclined to submit AI-generated content.
Don’t be afraid to:
- Quiz students on their own work. Quizzing students on their own paper is a tactic used in writing programs when students are suspected of submitting plagiarized or paid-for papers. And though faculty members may not like to use this method, it works.
- Use Blue Books and notebooks for handwritten, in-class assignments. In-class assignments or exams that require handwritten responses remain a surefire way.
- Update and Reinforce Your Plagiarism Policy. While you’re building your AI use policy, it’s also crucial to revisit your plagiarism guidelines. Make sure they’re updated to include AI-generated text, and that the consequences of breaching these rules are clear to your students. ECC’s Academic Honesty Policy has been updated to address this concern.
- Equip Your Students with AI-Identification Skills. As educators, it’s important for us to be able to spot AI-generated content, and the same goes for our students. Consider integrating a session into your syllabus to help students understand the subtleties that differentiate human-written text from AI-produced content.
- Discuss the Limitations of AI. As we adapt to this new AI-centric world, let’s remind our students that AI tools, while powerful, come with their own set of limitations. They are unable to replicate the depth of understanding, context sensitivity, and complex thought processes innate to humans. Highlighting these limitations will underscore the value of their individual contributions and diminish the appeal of AI-generated submissions.
- Use the Testing Center for tests (2 exams per course per semester). This is just a reminder that the ECC Testing Center can proctor 2 exams per course per semester. Contact the Testing Center (Union or Rolla) for guidelines.
- Consider adopting activities that include AI text generation as a sample, starting point, and/or opportunity for critique-and-revision. ChatGPT is interesting! It’s a valuable tool that all our students will have to use in their lives and careers. So engage with the tool and discover with your students what it can and cannot do.
But remember:
- ChatGPT can level the playing field for students with, for example, a poor writing education or a disability. Don’t be afraid to encourage usage while adding instruction on how to maximize the value of the software (and how to use it ethically).
- DO NOT rely on AI-text detection software to catch student usage. These tools are notoriously unreliable, providing large numbers of false positives and false negatives (often over 20% of the time, depending on the detection tool, the genre of text, and the AI text-generation tool used). They are also discriminatory against students who don’t natively speak English, giving false positives for these students up to 70% of the time (!).
- Actions that over-prioritize “catching/preventing cheating” can place undue stresses on equity, accessibility, and innovative thinking:
- Requiring handwritten submissions places additional burdens on students, especially those with some physical disabilities
- Adopting intensive surveillance tools or processes is likely to produce few gains while limiting students’ thinking and introducing additional biases and inequities into the learning process
Sources:
https://stearnscenter.gmu.edu/knowledge-center/ai-text-generators/
https://www.montclair.edu/faculty-excellence/practical-responses-to-chat-gpt/