Strategic Initiative Fund (SIF)

Primary Impact Area

The Michigan Deepfake Resilience Hub

The Michigan Deepfake Resilience Hub (M-DRH) is a tri-campus University of Michigan initiative that combines explainable AI, human-centered design, and education to help citizens, educators, and public institutions detect and understand AI-generated contents and protect themselves against AI-generated fraud. The project addresses a widening gap between the spread of AI-generated misinformation and the public’s ability to detect it, which can undermine trust in government, courts, education, elections, and public discourse, as well as a gap in existing detection tools that are often too technical for everyday users.

This M-DRH defines a unique solution by defining three key pillars: person-first design, education, and advanced technology. Its detection team has established a forensic foundation that unifies visual and audio analysis to expose synthetic manipulation across all media formats, while its human-computer interaction component bridges the human-trust gap by designing interactive interfaces that let users interrogate forensic evidence and make an informed judgment. The education and language-technology pillars translate these technical breakthroughs into modernized curricula and linguistic training to build long-term societal resilience.

By integrating advanced multimodal deepfake detection, interactive forensic tools, and modernized curricula, the hub aims to deliver a public verification portal for checking suspicious media, train more than 1,000 "Digital Integrity Ambassadors," and produce a response playbook and policy guidance to strengthen Michigan's resilience against AI-generated frauds.

U-M Teams

UM-Flint, College of Innovation and Technology (opens in new tab) (primary), UM-Dearborn College of Engineering & Computer Science (opens in new tab), School of Information (opens in new tab), College of Engineering (opens in new tab)

Secondary Contacts

Rada Mihalcea (College of Engineering), Hafiz Malik (UM-Dearborn College of Engineering and Computer Science), Cliff Lampe (School of Information)