In the digital purgatory of neural networks, a disturbing trend is emerging from Russia’s war in Ukraine. Families are using AI tools to animate the dead, creating eerily lifelike avatars of soldiers killed in action. The practice, condemned by Kyiv and human rights groups, lays bare a grim intersection of grief, technology and state propaganda.
The phenomenon first surfaced on social media platforms like VK and Telegram, where bereaved relatives post videos of AI-generated ‘ghosts’. Using off-the-shelf deepfake software or custom algorithms, they overlay a deceased soldier’s face onto a living body, sometimes syncing speech patterns or even generating new messages. The results are uncanny: a pixelated smile, a synthetic voice saying ‘I love you’, a digital hint that death is merely a system error.
But critics argue the Kremlin is exploiting this technology to spin a noble narrative around its ‘special military operation’. By encouraging these digital resurrections, Moscow fosters a culture of martyrdom, sanitising the true cost of a war that has claimed tens of thousands of lives. ‘It’s a new form of psychological warfare,’ says Dr. Anya Petrova, a Moscow-based digital anthropologist. ‘The state co-opts private grief, turning it into a tool for mobilisation. Each AI avatar becomes a recruitment poster.’
Ukrainian officials have called for international condemnation. ‘This is necromancy dressed as innovation,’ said Mykola Zadorozhnyi, a member of Ukraine’s Cyber Security Council. ‘Russia cannot bury its dead, so it digitises them. But you cannot resurrect a lie.’ The practice raises profound ethical questions about consent and dignity. Did these soldiers ever agree to have their likenesses immortalised? Can an AI truly honour a human life, or does it merely reduce it to data points?
From a technical lens, the quality varies. Most videos are crude, with flickering eye movements and robotic intonation. But as generative AI improves, the line between real and synthetic blurs. Experts warn that within a year, these avatars could be indistinguishable from live footage. ‘We are sleepwalking into a world where the dead can speak to us on demand,’ says Julian Vane, Technology & Innovation Lead. ‘It’s the ultimate user experience of grief, but at the cost of our shared reality. This isn’t about healing; it’s about manipulation.’
Silicon Valley built these tools for entertainment, not for weaponising loss. Yet here we are, watching families choose between truth and comfort. The Kremlin denies any official role, but independent investigators have traced state funding to companies developing ‘memory-preservation’ algorithms. The same technology could soon be used to resurrect politicians, revolutionaries or dissidents, for good or ill.
In one particularly chilling case, a mother named Yelena shared an AI video of her son, a 22-year-old conscript, wishing her a happy birthday. ‘I know it’s not really him,’ she wrote, ‘but it helps me sleep at night.’ This is the human cost of our tech-enabled age. We have the power to build worlds, but we have not yet learned to respect the boundary between life and algorithm.
Regulators are scrambling. The EU’s Digital Services Act could be used to ban such content, but enforcement is patchy. A broader conversation about digital sovereignty and the right to be forgotten is overdue. As Julian Vane puts it, ‘We need an ethical kill switch before we lose control of memory itself.’
For now, these AI ghosts haunt the edges of the war, a reminder that every new tool contains the seeds of its own misuse. The dead deserve more than a simulation. And the living deserve the truth, however painful.
