Selected as one of two recipients from the conference's inaugural proceedings, Fredrikson's paper took the then-recent discovery that AI models could be fooled by small, carefully crafted changes to their inputs and asked what it meant for building secure systems around them. The researchers created a framework for modeling the threats these systems faced and introduced new techniques for crafting adversarial inputs, demonstrating how even changes so small they were often imperceptible to people could cause a model to make the wrong prediction. Their work helped establish the field of adversarial machine learning, which studies how AI systems can be attacked and how to make them more secure.

The award-winning research was led by Nicholas Papernot, then a doctoral student at Pennsylvania State University, who worked with Fredrikson not long after the latter joined CMU as a faculty member in the Software and Societal Systems Department. The paper's co-authors included Patrick McDaniel, then a professor at Penn State; Somesh Jha, a professor at the University of Wisconsin–Madison; and Ananthram Swami and Richard Harang, researchers at the U.S. Army Research Laboratory.

Fredrikson and his co-authors were recognized during the EuroS&P 2026 conference earlier this month in Lisbon, Portugal.

For more information, visit the EuroS&P website.