The May 29 breakout session, Navigating the AI Code Jungle: Pros and Cons of AI-Assisted Software Development, is the brainchild of colleagues from Cornell University Library, Student and Campus Life, and the College of Veterinary Medicine.

This guided journey through tangled vines of code and unexpected traps for developers is led by Phil Robinson (Library IT) and features lightning-round presentations by Sarah Chintomby, Shinwoo Kim, and Matt Connolly from Library IT, Scott Ross (CVM IT), and Fermin Romero and Phil Williammee from Student Services IT.

“We are definitely experiencing another revolution in how software is designed, built, and enhanced,” said Phil Robinson, Cornell University Library Director of Software Development. “Cornell IT groups are using AI in diverse and fascinating ways. One important direction is the use of AI tools to assist with software development, process automation, testing, and learning.”

Robinson said, “In the Library IT group, for example, we have been leveraging the AI tool GitHub Copilot to improve skills in languages like Python and Ruby on Rails, and to improve testing and troubleshooting. Librarians have also been interested in applying these tools in their academic teaching and outreach work with students and faculty.”

Across campus in the Office of Student and Campus Life, the team led by Infrastructure and Application Development Manager Fermin Romero, along with his teammate Phil Williammee, has developed a different proof of concept. Romero explained, "We're exploring AI applications that might help with data access by enabling natural language querying for non-technical users. The idea is to bridge human language and structured data."

An important animal health application is being built with the help of AI in the College of Veterinary Medicine. Assistant Director of Application Development and Integration Scott Ross said the app’s users will need to “adopt a growth mindset and design thinking” to maximum the benefits of the new tools and Romero agreed, adding that a collaborative mindset is also helpful. “Think of AI as a partner, not a replacement for developers,” he said.

In addition to design thinking and testing rather than trusting the results of her generative AI tools, Sarah Chintomby, a Library Applications Developer, has learned to increase the accuracy of those outcomes by improving her prompts. She said, “Provide context within your prompt and be specific about language and version compatibility and accessibility. I also would recommend asking the generative AI chatbot follow-up questions about alternative solutions.”

Chatbots, Bears, and More

In a different breakout session, Associate Professor of Earth and Atmospheric Sciences Toby Ault and climate researcher Marty Sullivan will lead participants through their experiences using Chatbots and large language models (LLMs) to promote learning.

Symposium attendees can also learn to generate images like a bear running through the city and other fun graphics in the afternoon promptathon. The example below was created in Microsoft Copilot/Bing by CIT Manager, Cloud Operation & Service Management Sean Walsh.

Poster sessions provide further opportunities to listen, learn, and ask questions about a variety of AI-inspired use cases. From Weill Cornell Medicine, Director of Educational Computing Doug Cohen will be sharing his perspectives of AI in Education in an online poster presentation. In Statler Hall, Director of the Cornell Center for Advanced Computing Rich Knepper is presenting a poster on Empowering Researchers to Tackle Complex AI/ML and Data-Intensive Computing Challenges.

Watch the event page for more poster and breakout session updates.

Register Soon

Registration for the event indicates the topic of AI in Higher Education is of great interest to instructors, researchers, and students although staff are currently the largest group represented. Planning to attend in person? Register today! Due to Statler Hall catering deadlines, registration to attend in person closes at 11:59pm on Friday, May 17.