5 An Introduction to AI in an Honor’s Course (Chemistry)

Description of resource(s):

The resources include Powerpoint Slides that were used for a discussion-based introduction to Artificial Intelligence. These resources can be re-used by other faculty members who seek to introduce their students to AI. This was a 1-credit pilot course developed for an interdisciplinary honors program at UT Austin. The initial course can be updated or expanded based on the needs of students and the advancement of AI technologies.

Links to Resources:

Why I implemented this:

The resources were prepared for an undergraduate course to provide a discussion-based introduction to Artificial Intelligence (AI), focusing on its applications, recent developments, and ethical implications. Geared towards first and second-year undergrad students in CNS, the course included short weekly lectures, each followed by a discussion session. The group-based discussion involved students getting together in smaller groups of 3-4 participants, discussing specific ideas, and then reporting out to the rest of the class. Once all groups had reported, then each person could add additional comments based on other groups’ reports. The lectures were prepared to encourage critical thinking and engagement with the material. The curriculum covered foundational AI concepts such as neural networks, and explored current trends, and applications in various fields, and delves into the ethical considerations and societal impacts of AI technologies. Students became more familiar with the use of large language-models (LLMs) through course discussion sessions. For example, students would share prompts, the LLM responses, and their thoughts on how the model performed across different tasks. Each group worked independently, and then a “note taker” shared the response with the rest of the class.

My main takeaways:

The main takeaway from this course is that students are eager to discuss AI, and it is important to allow the students to explore AI technologies within the context of their own interest. Each student has their own interests, whether academic interests relating to their undergraduate research thesis, or personal interests such as video gaming. For this reason, it is important to make the class flexible rather than having an overly-prescriptive set of topics. End-of-semester presentations were very helpful since it allowed the students to explore specific AI contexts in more depth. For a successful course, the contents topics should be somewhat fluid. In addition, AI is evolving at a rapid pace, and the instructor should keep up with new developments, such as the launch of new models, or new manuscripts. These can be incorporated into the course throughout the semester.

What else should I consider?

The field of AI is moving at a rapid pace, it is likely that the slides will become outdated. Instructors should use this as a starting point but should do their own literature search and provide up-to-date material for students. In particular, Context is also important, as stated above, students have their own interests, and adapting the topics to the students interests can be important for engagement.

One pitfall is that the course was pass-or-fail, and attendance was not as regular as standard exam-based graded courses. For this reason, instructors may consider making attendance mandatory, or adapting the course to make each week’s class more “modular” so that students who miss class can pick up again the following week.

In terms of the audience, the course was targeted at students in the CNS Dean’s Scholars program at UT Austin. Most of the students were either Biology, Chemistry, Physics, or Neuroscience majors. The course could be adapted for other students to cover a broader range of topics that may appeal to other majors. However, the content should not be too broad given that students prefer topics that align with their interest.