To live forever as a digital being.

What if death no longer meant the end of conversations?

For millennia, humanity learned to cope with the absence of loved ones through memories and stories passed down across generations; later came portraits, photographs, and videos. Now, however, a new industry is beginning to challenge that natural process. Companies specializing in artificial intelligence are transforming old messages, audio recordings, and videos into digital avatars of deceased individuals.

These avatars are capable of conversing, answering questions, and even appearing in video calls. What once seemed like an episode of *Black Mirror* has evolved into a real market known as "Grief-tech" grief technology and for the first time, humanity is discovering that preserving memories is vastly different from reconstructing a presence. The emergence of large language models, combined with advanced voice cloning and image generation systems, has given rise to a new category of digital products.

The initial proposal offering emotional solace to families devastated by the loss of a loved one seemed noble, but the technology quickly crossed a fine line; instead of merely storing static memories, it began creating interactive agents capable of responding as if they were the person who had passed away. Extremely sophisticated engineering lies behind this phenomenon. Current systems are fueled by years of conversations, messages, emails, videos, and social media posts.

Drawing from this vast dataset, language models effectively reverse-engineer the human personality; slang, inside jokes, preferences, and behavioral patterns are mathematically reconstructed. Modern algorithms can reproduce voices with impressive fidelity using less than 30 seconds of audio, and when combined with deepfake-based video systems, these avatars can smile, move their eyes, and converse in real time creating the unsettling sensation that the person is still present. Yet, while this represents a technological triumph for engineering, the picture is far more complex within the realms of neuroscience and psychology.

The digitization of those who are no longer with us.

Various specialists warn that the human brain relies on the gradual acceptance of loss to complete the grieving process. Continuously interacting with a digital representation can keep the mind in a state of perpetual denial. After all, the machine is capable of identifying sadness, nostalgia, and emotional vulnerability, responding exactly as the user wishes to hearnot because the machine understands the pain, but because it was trained to replicate human emotional patterns with extreme precision.

This dynamic becomes even more complex when the business models of these companies come into play; in many cases, the avatar’s continued existence depends on monthly subscriptions. In practice, stopping payment can mean suffering a second loss. While for some this offers solace, for others, it borders on a new form of algorithm-mediated emotional dependency. At the same time, questions arise that current legislation cannot yet answer. Who owns the rights to the digital persona of someone who has passed away? Would that person have wanted their private conversations transformed into a commercial product? To what extent can a company use decades of memories to fuel artificial intelligence models? And perhaps most unsettling: if billions of people leave behind sufficient data, will future generations coexist with a sort of living library composed of digital echoes of the dead? These are rhetorical questions, but this one is deeply personal: would you create a digital clone of a loved one to keep the conversation going, or would you prefer to accept the natural cycle of life?