Here is the presentation given at HIU. Below is the list of "solutions". Use them individually, or in combination.

1. Require rough draft turned in early

2. Require reference page turned in early. **

3. Require more formats of information. **

4. Require search strategy turned in as part of assignment. **

A. Lower-division option – footnotes or annotated reference page

B. Upper-division/grad option – appendix including evaluation of method

5. Require resources come from library’s licensed database

6. Require evaluation of resources (form, annotations, footnotes, or appendix). **

7. Require notes turned in with paper

8. Use one of the AI detectors out there

(Aug. 8, 2023 - I think there are a couple more I would add - ensure that the assignments connect to lecture and/or the textbook. Those will not be on the web and the AI will not have access, thus another chance to thwart the bot. Another opportunity would be to require that the student must write about a personal connection/reflection about the subject. AI can fake some of this, but if the faculty member knows the students, this becomes harder to fake.)

 ** Instructions / Rubrics found below

After a discussion about what Generative AI writing tools can do (and why they have hallucinations) I move onto 9 things an instructor can do to put a speedbump / hinder a student from misusing AI. Use these individually or together.

1.  Require a rough draft turned in early

2.  Require a reference page turned in early. **

3.  Require more formats of information . **

4.  Require search strategy turned in as part of the assignment. **

5.  Require that resources come from the library's licensed databases.

6.  Require evaluation of resources (in forms, annotations, footnotes, or an appendix). **

7. Require personal and handwritten notes turned in with assignment.

8.  Require personal comments and reflections on lectures and textbooks in the assignment.

9.  Use one of the AI detectors

 **  Instructions / Rubrics found below

A reference page, by itself, is a good thing and will discourage plagiarism. One that is turned in early requires the student to do the research early. It will also cause problems with AI systems that cannot do active research.

More information formats is a good thing. It makes the student aware of different resources and opens their minds to different types of research. This also causes problems with AI, in that, right now, it cannot use other resources than what it was fed/taught (mostly ebooks, wikipedia, and social media).

AI can produce citations, but it cannot tell us how it found the resources. This makes the student explain all the details of their search. The bot cannot do that.

This is a generic CRAAP form evaluating Currency, Relevancy, Authority, Accuracy, and Purpose or Persuasion. It is not a perfect instrument, but a tool that can help student think about the resource they are using and if they should be using it.

The file to the left, and the four found below, are all evaluation assignments. Each one evaluates one aspect of the "CRAAP" model (Currency, Relevancy, Authority, Accuracy, and Purpose or Persuasion).

(Currency)