In a deeply unsettling development, families of Russian soldiers killed in Ukraine have begun using artificial intelligence to create digital avatars of the deceased. The practice, which involves feeding photographs, videos, and voice recordings into generative AI models, produces eerily lifelike chatbots that mimic the fallen soldier’s speech patterns and mannerisms. Some families even report holding full conversations, receiving birthday reminders, and hearing ‘I love you’ from the dead. British technology ethics authorities have condemned the practice, warning of severe psychological harms and the wider societal dangers of AI-driven grief exploitation.
The technology behind these ‘digital resurrections’ is not new. Deep learning models such as GPT variants and synthetic voice generators have been commercially available for years. What is new is their deployment in a war context, where families are desperate for closure. Russia’s state media has reported on several cases, framing them as patriotic memorialisation. However, the British AI Ethics Board has labelled the trend a ‘black mirror nightmare’, pointing to addiction, fraudulent emotional dependency, and the erosion of natural mourning processes. ‘Grief is not a problem to be solved by algorithms,’ said Dr. Eleanor Shaw, the board’s chair. ‘These digital ghosts offer a hollow comfort that ultimately prolongs suffering.’
The technical process is straightforward but morally fraught. Families upload personal data to a platform, which then uses speech synthesis and natural language processing to generate responses. The avatar lives in an app, calling the user or texting them on anniversaries. One Russian widow described her avatar as ‘my Sasha, still here’. Psychologists warn that such substitutes can prevent the brain from accepting loss, leading to complicated grief disorder. Moreover, the data remains stored on private servers, raising digital sovereignty issues about who controls these memories and for what purpose.
This is not the first instance of AI grief tech. A US startup offered similar ‘deadbot’ services in 2021, but quickly shut down after public outcry. In China, some people use AI to ‘speak’ with deceased relatives during Qingming festival. Yet the Russian military context amplifies every concern. There are reports of children being given access to these avatars, blurring their understanding of death. The British government has called for an international treaty banning AI-driven impersonation of the dead without prior consent. The Ministry of Defence is also monitoring potential use in propaganda: imagine fake soldier avatars urging others to fight.
For now, the genie is out of the bottle. The AI models are open source, and the ethical board admits regulation is difficult. But the human cost is clear: we are trading genuine mourning for digital palliatives. As one British ethicist put it: ‘We are building a world where the dead never truly leave — and that might be the most dystopian future of all.’
