Text-matching software, such as Turnitin or Ouriginal, shows the percentage of a submission that matches other sources. This typically will not identify AI-generated text because it is original text, just as it does not identify work written by contract cheating providers for student customers. However, this is rapidly changing as companies are racing to generate the newest way to identify AI-generate text.

There are already new tools available that claim to identify AI-generated content. These may be open-source tools that may become unavailable at any time, or proprietary tools that come with a cost. These tools are also unproven and best used as a guideline and not as definitive proof that a given passage of text was generated by AI.

Of course, the development of AI-detection is going to create an arms race of finding ways to outsmart that identification, as well. An instructor’s academic judgement in knowing what is or isn’t their student’s writing remains the best way to identify potential academic misconduct.