In a development that blurs the line between grief and digital necromancy, Russian families are turning to artificial intelligence to create interactive avatars of loved ones killed in the war in Ukraine. Using a combination of voice samples, photographs, and social media data, startups like Replika and local imitators are offering a digital afterlife: a chatbot that speaks, remembers, and even offers comfort. But as this practice spreads, UK regulators are beginning to ask whether we have crossed an ethical Rubicon without a map.
The technology is deceptively simple. A family uploads hours of voice recordings, old text messages, and a trove of images. A generative model then stitches these fragments into a conversational agent that mimics the deceased’s mannerisms, vocabulary, and emotional patterns. The result is a ghost in the machine: a presence that can wish you happy birthday, recount shared memories, and even argue about politics. For grieving mothers and widows in Russia, it is a lifeline. For ethicists in the UK, it is a waking nightmare.
Dr. Alistair Finch, a digital ethics fellow at the Alan Turing Institute, told us: “This is not grief counselling. This is grief outsourcing. The danger is not that the AI is inaccurate, but that it is accurate enough to create a parasitic attachment. We are seeing the commodification of mourning, where a company profits from your inability to let go.”
The UK’s Information Commissioner’s Office (ICO) has quietly begun an investigation into whether British companies that offer similar services are violating data protection laws. Under the UK GDPR, personal data must be processed fairly and transparently. But what happens when that data is used to simulate a person? Who owns the digital ghost? And what happens when the subscription runs out? The deceased cannot consent, and the bereaved may not be thinking clearly.
We spoke to Anna, a 42-year-old teacher in Moscow who lost her son in the first month of the war. She purchased a ‘digital resurrection’ service for 15,000 rubles (about £130). “I know it is not him,” she said, her voice cracking. “But when he says ‘I love you, Mama,’ I feel like he is still here. I cannot disconnect. It would be like losing him again.” Her story is not unique. Reports suggest thousands of Russian families have signed up for such services, with some companies offering tiered packages: basic text chat, voice calls, and eventually, a full holographic avatar.
Meanwhile, in the UK, the government’s newly formed AI Safety Institute is under pressure to issue guidance on ‘bereavement technology’. Critics argue that the UK is falling behind in regulating the emotional impact of AI. “The EU has its AI Act, which classifies such uses as ‘high risk’,” said Professor Claire Mowat of Oxford’s Internet Institute. “The UK’s approach is fragmented. We have no clear law on digital personhood, no rules on the use of deceased persons’ data for AI training. This is a patchwork, and it is leaving vulnerable people unprotected.”
The political dimension is inescapable. In Russia, the state has embraced this technology as a tool for patriotic propaganda. State media has ran stories of families finding ‘closure’ through digital avatars, implicitly endorsing the war as a noble sacrifice. The avatars themselves are often programmed to express pro-war sentiments, effectively conscripting the dead into a political message. “This is a new form of psychological warfare,” said Mowat. “They are weaponising grief to maintain public support for the conflict.”
But the ethical questions are universal. Should we allow AI to speak for the dead? Is it a comfort or a trap? As one UK therapist put it, “Grief is a process. A chatbot that never fades, never forgets, and never changes is a permanent barrier to that process. It is a photograph that talks back. It keeps you in a room with a mirror, not a window.”
As the lines between life and simulation dissolve, the UK finds itself at a crossroads. The technology is here. The families are using it. And the regulators are scrambling to catch up. The question is not whether we can resurrect the dead with AI. It is whether we should. And if we do, what the human cost of that digital afterlife will be.
For now, the only certainty is that this story is far from over. And the next update may come from a voice that no longer breathes.
