The project Overlife will follow several key phases, each giving birth to a new generation of AI:

• Gen 1: 24/7 learning using personal data like photos, videos and recordings. This is the easiest but the longest way to obtain enough material for an AI to “learn” someone and be able to sound, react and look like that person. It will eventually be completed by dedicated sessions to obtain missing or insufficient content.
• Gen 2: Template, customized with criteria and experience. The biggest challenge is to determine the basics of humanity, to be able to rebuild a human, like we recreate all colors by mixing only red, green/yellow and blue. That template will be configured with specific attributes like gender, physiology, culture, age, etc. Then a smaller amount of personal data will be necessary to refine the personality of the “Vuman”.
• Gen 3: This one will mix both previous generations and add age evolution, for parents who lost a child or young adults who lost a partner. It will give them the opportunity to see their lost ones virtually “grow up” and age with them.
• Gen 4: Why limit to humans? Pets are members of the family too and losing one can hurt as much as losing a brother or a sister. This generation will be dedicated to reproduce their behaviour and interactions.
• Gen 5: Scalable autonomous Vuman. Couples who can’t have children suffer too. Based on a Gen 3 but with a mix of both parents’ attributes.
• At this time, it will become possible to rebuild extinct species or simulate evolution. This generation would not only be interesting for dinosaurs lovers but also for scientific research.
• Gen 7: Self-aware, independant and autonomous beings, like organic species.

Note: International laws will be necessary to regulate the virtual life in order to avoid abuse, even if AI are not self-conscious yet, because coming AI projects will be built on personal data and attributes.