BANNER

Bio 162: Ichthyology (The Study of Fishes)

Generative AI, Pattern learning, Understanding

Reflection on Instructional Change:

My activities most aligned with Empowering Students Using AI. I want students to understand what Generative AI is, how it works, the limitations and the consequences. So, prior to class, I asked every student to get an account at OpenAI and to try to use ChatGPT for a few minutes. Then, they were to write a paragraph answering the prompt "How do you think ChatGPT comes up with its answers, i.e., how does ChatGPT work?" I wanted their perceptions. In class, I asked students to share their answers. The most common answer was that ChatGPT examines a large body of information (e.g., the Internet), seeks out a good answer and reports it. I then explained how machine learning (pattern matching) works using John Conway's Game of Life, an example of Cellular Automata. It shows the appearance of "intention" which is not actually there. I then explained exactly how Gen AI works and that it is really just pattern matching, not understanding. Finally, we talked about how this affects use and misuse of Gen AI in a science context

The students were surprised to learn how ChatGPT actually works and we had good discussion about what this means in terms of what you can use ChatGPT for, and what you should not trust ChatGPT for. Students were also surprised how well a system which actually does not understand what it is saying can appear to understand what it is saying.

That said, ChatGPT is not all negative, it just has to be used with great caution.

Evidence of Instructional Change

evidence of instructional change
Some interesting sample patterns from Conway's Game of Life. This game shows that "life-like" behavior can result from a set of very simple rules, i.e., something can appear to be alive or "thinking" without actually doing so, e.g., ChatGPT	Stephen Wolfram's book "What is ChatGPT Doing?" is an excellent introduction to exactly how generative AI works.	Take home message

Practice and Reflect Teaching and Learning with AI:

The course forced me to really dig into AI in a way that I had been meaning to do, but did not have the time. I now follow the AI literature, as presented in such journals as Science and Nature, much more closely, as well as following the weekly developments on how AI is being used and misused in academic (often discussed in the newsletter Inside Higher Ed).

Course Syllabus

From Canvas:

Ronald Coleman (He/Him/His)

Author
Teacher

Created Sep 10 11:32am | Posted Sep 10 11:32am

Discussion Topic: Bio 162: ChatGPT

Hi folks,

1. As I mentioned yesterday in class, I need each of your to go to OpenAI.com and sign up for a free ChatGPT account.

2. I then want you to play with ChatGPT for at least 10 minutes. Try some things. Get a feel for it.

3. After having done that, I want you to answer one question (email it to me) with the subject: "Bio 162: How ChatGPT works. "

The question is "How do you THINK ChatGPT works, i.e., how does it come up with its answers?"

I do not want you to investigate how ChatGPT works at this point. I am interested in your PERCEPTION of how it works based on what you may have heard about "AI" (Artificial Intelligence) and your brief experience using it. Your answer should be a few sentences long.

Email me this by next Monday morning (Sept 15).

Thanks.

Ron C.

rcoleman@csus.edu

Faculty Bio

faculty bio
	Ron Coleman, Professor of Biological Sciences

Funding
	This work was supported by: US Department of Education Hispanic Serving Institution Project TECH4Equity US Department of Education Hispanic Serving Institution Project STEM4Equity P031C210012 *Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation or the US Department of Education.