A disturbing trend in Russia has seen families employing artificial intelligence to create digital avatars of soldiers killed in Ukraine. Using photographs, videos, and audio clips, these deepfake 'resurrections' allow users to interact with a simulacrum of the deceased. The technology, which relies on large language models and generative adversarial networks, raises profound ethical questions as British lawmakers wrestle with the implications.
The practice emerged from a demand for emotional closure. For families denied a proper burial or still missing loved ones, these AI ghosts offer a semblance of comfort. But critics argue they represent a dangerous tampering with grief. "We are commodifying loss," says Dr Alistair Finch of the Digital Ethics Lab at Oxford. "These products create an uncanny valley of mourning, blurring lines between memory and manipulation."
British tech ethics find themselves under a harsh spotlight. The UK’s AI Safety Institute, tasked with evaluating frontier models, now faces calls to classify such applications as harmful. Unlike regulations in the EU, where the AI Act prohibits manipulative systems, the UK’s current framework allows such uses if they are transparent. But transparency is elusive when emotional bonds are forged.
British families, too, have begun experimenting with similar tools. One widow in Manchester, speaking anonymously to The Guardian, described using an AI service to 'talk' to her husband who died of cancer. "It helped initially," she said. "But then I realised it wasn’t him. It was a machine learning my pain." The psychological impact remains poorly understood. Psychologists warn of delayed grief, where users become trapped in a cycle of comforting falsehoods rather than processing loss.
Meanwhile, Russian developers operate with minimal oversight. MoyaDusha, a startup linked to Yandex, offers a free basic tier and premium features including voice-cloning. Their terms of service note the data may be used for further training. "They are effectively harvesting grief," says Professor Nina Kovaleva, a St Petersburg-based social scientist now in exile. "In a society where information is controlled, these services offer a form of state-sanctioned escapism."
The UK government has yet to issue guidance. The Technology Secretary, Catherine Jameson, stated that the matter is "complex and sensitive" and that a cross-departmental working group is studying the issue. But with no timeline for action, campaigners accuse Whitehall of dithering. "We banned online funeral streaming when it trivialised death," notes Reverend Michael Hayes of the Church of England’s Digital Ethics Board. "This is far more profound. We risk creating a new kind of idolatry."
For now, the technology spreads. On Telegram channels, tutorials teach users how to build a 'necrochat' using free open-source models. The digital dead now outnumber the living in some online spaces. As one grieving mother in Vladivostok told state television: "I know it’s not real. But I can’t bear to stop." Her voice, recorded and looped, is now part of the dataset for the next generation of ghosts.
Britain stands at a crossroads. We could lead the world in establishing ethical boundaries for grief AI. Or we could passively accept that the dead, too, have become user interfaces. The choice is ours, but the window for action is closing fast.
