Aftertalks #2 - Will AI Commercialize Human Clones After Death? | AI-swers

What if you received a perfectly realistic message from someone who died years ago? In this AI-swers Aftertalks episode, we examine how ChatGPT, Claude, Gemini, DeepSeek, Llama, and Grok approached digital afterlife technology: AI-generated voices, interactive avatars, personality simulations, and virtual recreations of deceased people. Modern systems can combine voice cloning, facial animation, large language models, vector databases, and Retrieval-Augmented Generation to create an interactive simulation built from a person’s messages, recordings, memories, and digital footprint. But simulation is not resurrection. The system does not remember, feel, grieve, or understand. It predicts how the deceased person might respond—and an incorrect response could cause profound psychological harm. Could these avatars provide closure when used within a structured therapeutic process? Or would commercial platforms exploit grief, create emotional dependency, and prevent people from accepting loss? CHAPTERS 00:00 A message from someone who died 02:00 Digital resurrection is already becoming possible 03:20 The Super Brain and interactive avatars 04:00 What exactly is being recreated? 05:00 The six-model digital-afterlife debate 06:00 How modern voice cloning works 08:00 Simulating personality with LLMs and RAG 10:00 Why the avatar does not understand you 11:20 How a digital person feels psychologically real 12:00 The uncanny valley and simulation errors 14:00 Experiencing loss for a second time 14:40 Why some AI models cautiously supported it 16:00 Complicated grief and the search for closure 17:00 AI avatars as proxy therapy 18:00 Could a controlled step-down process help? 19:20 Digital memory, consent and the final boundary