Four faculty from Miami University College of Engineering and Computing presented at Miami University's AI Symposium on November 1, including Suman Bhunia, Samer Khamaiseh, DJ Rao, and James Walden. Suman Bhunia and DJ Rao's presentation was titled “With ChatGPT Do We Have to Rewrite our Learning Objectives? Case Study in Cybersecurity.” Their presentation focused on the impact of AI tools like ChatGPT and GitHub CoPilot on what and how to teach cybersecurity courses in the future. They examined how learning outcomes would need to move to higher levels in Bloom’s taxonomy,  since AI tools can automate many lower level learning tasks.

Samer Khamaiseh and James Walden presented on the security risks of AI models being integrated into the software we use. They examined how AI models are susceptible to a variety of attacks, including adversarial examples, which can fool an AI to misclassify a stop sign as a yield sign. Dr. Walden’s part of the presentation focused on the dangers of prompt injection to large language models (LLMs), showing how AIs can be tricked into performing malicious actions through hidden prompts embedded in web pages, images, and PDF documents.