Edward J. Delp

As artificial intelligence makes it increasingly easy to create realistic fake videos, images and audio, researchers are working to answer a critical question: How can we know what is real?

Edward J. Delp, the Charles William Harrison Distinguished Professor of Electrical and Computer Engineering at Purdue University, addressed that challenge as a keynote speaker at the inaugural International Symposium on Synthetic Media Attribution and Detection, or ISSMAD 2026.

The symposium, co-sponsored by Google and the IEEE Signal Processing Society, was held Aug. 20-21 at Google's Sunnyvale, California, campus. The event brought together researchers, industry experts and technology platform representatives to tackle the growing challenges surrounding AI-generated and manipulated content.

Delp's keynote, "Deep Fakes and Manipulated Media: The Agony, the Ecstasy, and the Extinction Event," explored both the opportunities created by advances in generative AI and the risks they pose when realistic synthetic media is used to deceive people.

“We have reached a point where creating convincing synthetic images, video and audio is becoming easier and more accessible,” Delp said. “That creates exciting possibilities, but it also creates serious challenges for knowing whether the content we see and hear is authentic. We need technologies that can not only detect manipulation but also help establish where digital content came from and how it was created.”

Delp has studied media forensics and manipulated content for decades. His research includes developing tools that use artificial intelligence, machine learning and signal processing to identify altered media and determine how it may have been manipulated. His work has examined deepfake videos as well as manipulated images, documents and other forms of digital content.

ISSMAD was designed to encourage collaboration across academia and industry at a time when rapidly improving generative AI tools are making media authenticity an increasingly important issue.