Note: This research is being conducted in Fall 2026 with Muhlenberg undergraduate student researcher, Joseph Tsocanos.

An evaluation harness for testing how vision-language models perceive and respond to images protected by anti-AI-training tools such as Nightshade and Glaze.