Deepfakes, Synthetic Audio and Conversational Agents as Teaching and Learning Tools

Generative Adversarial Networks (GANs) use advanced allgorithms to create synthetic audio and visual media. Neural networks in GANs can be used in ingenious ways to make deepfakes, which change people’s features and voices without a hitch, and shallowfakes, which add fake actions to videos, that can link little lies to partial truths. While it may be challenging for students to identify and combat deepfakes, the speaker emphasizes the importance of verifying the authenticity and ensuring the legitimacy of audiovisual recordings by gaining a deeper understanding of the differences between real and fake conversational actors and synthetic audiovisuals of bogus media.