The concept of AI-generated voices of deceased individuals involves using artificial intelligence algorithms to recreate the speech patterns and vocal characteristics of people who have passed away, allowing their voices to be simulated in various contexts.