Teaching and Generative AI

Please remember if you don't have concrete evidence that a student used an AI tool, it may not be appropriate to accuse them of plagiarism or academic misconduct. 

 AI detectors are NOT reliable and it is difficult to determine whether a student used generative AI to complete their assignments. It is not recommended to rely on AI detectors but to use multiple sources to identify improper use of AI.

A few ways can be to:

• Check inconsistencies in the student's writing style. Are they using technical language that they have not used in past assignments?   
• Verify links to sources.
• Meet with the student to discuss their idea and writing process for their work. Their answers to your questions can help inform your decision.
• Turnitin, the only campus approved plagiarism detection tool available for free to SJSU users. You can enable Turnitin for assignments on Canvas. Please note: There is strong evidence of AI detectors failing including Turnitin, to identify AI-generated content and flagging human-generated content as AI.
  • Submitting a student's intellectual property to an unapproved website or tool is an Information Security issue and violates University Policy S16-13 and the CSU Information Security Responsible Use Policy. It also may violate FERPA, as some third party websites claim ownership of any content submitted to them, including student work. 

Sources: OpenAI: Educator considerations for ChatGPT, Vanderbilt University: Guidance on AI detection and why were disabling, FAQ: ChatGPT in the Classroom

If you suspect that your student used ChatGPT inappropriately or outside your AI policy. We suggest to:    

• Follow SJSU's Academic Integrity Policy.
• Reach out to the Student Conduct and Ethical Development.
• Read about how to approach a student regarding potential AI misuse.
• Consider it as a learning experience for students rather than taking disciplinary action. Speak with the student directly and ask them to explain their work while also explaining the purpose of the assignment and the objectives of the class.
• Ask yourself whether you have clear guidelines about AI usage listed in your syllabus? Is your AI policy easily accessible to students?
  • It may not be appropriate to take punitive measures against students for using AI tools if expectations of usage were not outlined and clear beforehand.
  • If your policy allows the use of AI tools in certain contexts, but the student has used the tools in a way that is forbidden by the policy, consider opening up a conversation with the student to help you understand if your policy is clear and if the specific limits you impose are a good fit for the assignment.
• Reflect on the assignment and always use your best judgement:
  • Did the assignments require students to show their own analysis and understanding of a topic? The use of ChatGPT and other AI tools may not be appropriate.
  • Did the assignment require the student to summarize information from multiple sources? Did the student cite ChatGPT correctly or indicated their use of it on this assignment? The student may have appropriately used ChatGPT.

Source: ChatGPT in the Classroom