Welcome to the CISAAD repository. This site hosts resources for educators, researchers, policymakers, or anyone curious about audio deepfake detection.
Learn to Spot Deepfakes
Check out our series of short videos on Expert-defined Linguistic Features (EDLFs), which introduce five perceptual linguistic features that can help you spot fake or manipulated audio. Also visit Education > Educational Resources > General Education Resources for a repository of articles, podcasts, videos, and more created or assembled by our team that deal with audio deepfakes and their impacts on society.
Teach About Deepfakes
Start out at Education > Educational Resources > General Educational Resources for a repository of articles, podcasts, videos, and more created or assembled by our team that deal with audio deepfakes and their impacts on society. Also check out our series of short videos on Expert-defined Linguistic Features (EDLFs), which introduce five perceptual linguistic features that can help you spot fake or manipulated audio.
For K-12 educators, check out Education > Educational Resources > K-12 Educator Resources to find a tailored collection of multimedia resources for students at all levels.
For university instructors, visit Education > Educational Resources > Relevant Syllabi for a collection of examples of publicly available syllabi from college courses that deal with issues related to audio deepfakes and language technology.
If you teach Computer Science or ML courses, check out our Caselets, a set of three short problem-solving exercises that walk students through developing an automatic detector for deepfake audio and introduce the real-world potential harms of linguistic discrimination and computational bias.
Find a Dataset
Visit Research > Datasets or, for a specific language other than English or for multispeaker interactional audio, check Research > Research Literature > Linguistics.
Read Research Work
Visit Research > Research Literature > Linguistics for papers on audio deepfake detection in languages other than English, including signed languages, in dialogic interaction, and more. Also check out Research > Detection Methods and Generation Methods or the Tools page for an overview of different automatic methods for generating and detecting deepfakes, including not only audio but also video, image, and text content.
Know More About the Team
Check out our Publications page for an updated list of papers, presentations, as well as recorded talks and podcasts from members of our team. Visit Meet the Team to see who has been involved in this ongoing research effort, from faculty to graduate, undergraduate, and high school students.
Find Deepfake Regulation Policies
Visit our Policy page for a collection of recent policies in the United States that seek to regulate deepfake content.