Turnitin’s Artificial Intelligence (AI) writing detection is here, and with it comes a need for some essential conversations between educators and students about potential AI misuse. How courageous these conversations are depends on how much communication takes place prior to, during, and after assignments/assessments are made.
AI is a game changer for sure, but in many important ways, the conversations we have won’t change as much as it may seem initially. So right here, right now, the question that I need answered as a veteran secondary educator is, “What do we know to be true and helpful to educators and students?”
Take a deep breath and let’s begin to map out how to approach students with concerns over possible misuse of AI.Is there a way to detect AI writing?
Coupled with educator insight, Turnitin's AI writing detection feature can support the creation of original work by illuminating areas of possible concern. However, as these tools to detect AI writing gain traction, communication and transparency between the educator and the student are crucial in establishing the trust needed to hold potentially difficult conversations. While not always possible to avoid frustration or defensiveness, the following is a pivotal first step to avoiding a surprise that may feel like a "Gotcha!" moment to students.Step one: Learn about Turnitin’s AI writing detection feature and share with students how it will be applied to their work.
As educators, it is of utmost importance to understand what the AI writing indicator means. Since this score isn’t currently visible to students, an educator needs to understand what the score indicates and how it might be used to facilitate impactful and meaningful dialogue. This score shows the overall percentage of the document that AI writing tools may have generated. The educator can view the score and the underlined passages in the report and consider possible reasons for the use of AI. Keep in mind that the AI writing indicator is best used to inform the educator’s judgment, not to be the sole measure of academic integrity.
- Students may benefit from the Ethical AI use checklist, which is a good way to help them focus on techniques to avoid even the appearance of academic misconduct. Be ready to discuss what each of the before/during points on the checklist mean and emphasize the parts most relevant for a particular assignment. Remind them to reflect on the second part of the checklist and that they may be using it as a starting point for discussions during a writing conference later.
- It is also important to remember that there is no one “right” or “target” score with the AI writing indicator, just like with a Similarity Score. Educators working directly with students are in the lead here, as they are best positioned to make these kinds of nuanced determinations, based on their in-depth knowledge of the students and assignment. For example, were students using AI to draft an introduction as a whole class activity to help get past the blank page or screen? Educators will need to acknowledge these distinctions when viewing the AI writing report and then can suggest a particular course of action, such as citing the AI-generated text to avoid the appearance of academic dishonesty.
While the chance of a false positive is slight, instructors may find value in diving into AI conversations: Handling false positives for educators. This resource offers suggestions for how to handle such situations after assignments are submitted.What happens when AI is detected in an assignment?
When/if AI is flagged using the AI writing indicator, educators have already laid the groundwork by learning about Turnitin's AI writing detection feature and communicating to students how it will be used. Then, educators can take the next steps as they prepare for and conduct essential conversations:Step two: Gather artifacts when AI misuse is suspected.
It is important for educators to gather artifacts such as previous drafts that clearly demonstrate the student’s voice or writing style. This information aids the educator in determining if there is a pattern or a larger picture that helps to inform the dialogue and possible solutions. Once the artifacts are collected, instructors may want to consider these questions:
- Is the style of writing different from other assignments?
- Is the student’s voice different from other submissions?
- Has the student included references to classroom discussions or other, more personal references unable to be replicated by AI?
- Are citations present? Are they correct for the information they are intended to support?
The highlighted sections from the AI writing report provides concrete data points to support the educator in showing the student where the work appears to not be the student’s own original work. These highlighted passages point to the student’s action and not the student, and thus aids in avoiding a defensive response. Having the data readily available provides information to better determine next steps.Step three: Be solution- and student-focused.
Consider why or how the student may have misrepresented the work of a generative AI tool as their own. Considering why they may have done this will help to suggest solutions which may require compiling some information or resources before the actual conversation.
- Is a rewrite of certain sections needed? Of the work in its entirety?
- Will a handout or link to a site about citations suffice or is a tutorial needed?
- Are directions needed on how to make a direct quote? Is paraphrasing needed?
- Schedule periodic check-ins with a peer, tutor, or instructor.
While it’s important to have some solutions in mind, it’s equally important to allow for a conversation where the student input is encouraged and listened to before “solving” it for them. A collaborative conversation will often have a greater long-term impact.Step four: Begin a dialogue with the student that focuses on specific documentation from both the educator and the student.
Use discussion starters, specific artifacts, and resources to suggest a solution to the AI usage that was concerning, while also steering the conversation away from assumptions that might not be productive.
Keeping the focus on clear communication around the writing process and the specific passages which first created questions will both guide the discussion and help to avoid the emotions and defensiveness that sometimes can derail such conversations.
Encourage the student to explain the choices they made regarding the use of AI and to share the artifacts that would illuminate their process and help to explain their choices.
Educator documentation of questionable work
Student artifacts to support writing choices
Writing style appears inconsistent with student’s previous work
An outline of the work planned
Voice is inconsistent with student’s previous work or not evident at all
Multiple drafts or submissions within Feedback Studio
Submission is less personal or somewhat generic in tone
Drafts showing proof of revisions, either from peers or by the student
A lack of citations for some ideas
A list of sources used to complete the assignment
AI writing score
This conversation may lead to an opportunity for a tutorial on citation or student voice, for example, or the educator might need to aid the student in finding more acceptable sources to include. Or the conversation may lead to an understanding that the AI usage was acceptable. Open conversations will help educators understand why the student made the choices they did and to support their learning.
It is of utmost importance for educators to provide clear direction as to what the student needs to do before submitting the assignment again. As mentioned before, using the Ethical AI use checklist for students can also help provide guidance as they continue working on this assignment as well as for future work.
Whatever the solution(s)--whether it’s a full or partial redo of the work, providing citations, etc.--giving a timeline and arranging for another conference may be just the support needed for the student to implement the solutions decided collaboratively. Be explicit and clear about expectations and provide enough flexibility here that the student feels free to express their ideas about what support(s) they may need moving forward.
As my coworker often reminds her toddler: “We can do hard things." There is no doubt that the appearance of widespread access to generative AI tools is going to change the face of education–and already has. As educators, these conversations with students allow us to move forward proactively and with the goal of helping students to be engaged learners, to be stronger thinkers and writers, and to use all available resources to meet these goals. Take another deep breath. Continue moving forward. We can do this.