The risks of using AI detection tools currently outweigh the possible benefits
If you decide to prohibit the use of AI in your teaching, we do not recommend using AI-detection tools to identify plagiarism.
There have already been many well-documented cases of innocent students being accused of academic dishonesty due to the complexity of the technology.
Problems with AI Detection
There are several reasons for why AI-detection tools are problematic:
- AI detectors are not reliable and often result in false positives. Even OpenAI (who developed ChatGPT) has not been able to produce a tool that will reliably detect AI-generated writing and shut down its AI classifier in July 2023, citing a “low rate of accuracy.”
- They can be biased, particularly against non-native English writers.
- The rapid pace of development of these technologies makes it challenging for detection software to keep up.
Communication With Students
If you do decide to ban AI, it’s important that you have a conversation with students about your rationale so that they understand why you’ve made this choice. To not do so is a missed opportunity to help your students think critically about the technologies in their lives as well as the way their choice to use particular technologies impacts their learning process.
For more on this topic, see Communicating with Your Students around AI