A disturbing trend has emerged from the fog of the Ukraine war: Russian families are using artificial intelligence to create digital avatars of fallen soldiers, effectively ‘resurrecting’ the dead for ongoing interaction. This practice, which blends grief with generative AI, has prompted urgent warnings from UK tech ethicists about the risks of normalising digital necromancy.
The phenomenon, first reported by independent Russian media, involves feeding photographs, voice recordings, and social media data into AI models to produce chatbots and deepfake video likenesses of deceased servicemen. These digital revenants can then ‘converse’ with loved ones, often using phrases and mannerisms scraped from the deceased’s online footprint. Some families report spending hours talking to these avatars, finding temporary solace in their simulated presence.
But the ethical implications are vast and troubling. Julian Vane, Technology & Innovation Lead at a London-based think tank, describes this as “a Black Mirror episode playing out in real time”. He warns that while the impulse to console the grieving is understandable, the technology risks eroding the natural grieving process and creating a dangerous dependency on synthetic relationships. “We are seeing the commodification of loss,” Vane says. “These services profit from trauma, and there is no regulatory framework to ensure they operate with any duty of care.”
The UK’s digital ethics advisory board has been urged to scrutinise the practice, particularly as similar services begin to appear in the West. Vane points out that the same AI tools could be used for less benign purposes, such as spreading disinformation via fake voices of the dead or manipulating grieving families. “Once you normalise talking to dead people through a screen, you open the door to a ransomware attack on memories,” he warns.
From a technological standpoint, these ‘griefbots’ rely on large language models and voice cloning, techniques that have advanced rapidly in the past two years. Companies offering these services typically charge a fee, raising questions about exploitation. Vane notes that while some startups position themselves as providing therapeutic support, the lack of clinical oversight is alarming. “This is not therapy. It is an algorithm trained on loss, and it has no ethical compass.”
The case also highlights a broader issue: the digital sovereignty of the deceased. Who owns the data used to recreate a person? And should an AI have the right to simulate a consciousness without explicit consent from the individual before death? In Russia, legal loopholes allow companies to use deceased persons’ data without permission, as privacy laws do not extend beyond the grave. Vane argues that this is a regulatory vacuum that must be filled urgently. “We need digital estate laws that give people control over their posthumous identities. Otherwise, we risk a future where our dead are endlessly reanimated by algorithms we never agreed to.”
For now, the practice continues, driven by a combination of unprecedented grief and technological capability. Vane’s final thought is characteristically sharp: “Technology amplifies our humanity, both the best and the worst of it. This is not innovation. It is a failure of empathy dressed up in code.”
