Roman Mazurenko, a Belarussian-born tech entrepreneur, was run down by a speeding car on a Moscow street in November 2015. In his early 30s at the time, Mazurenko died from his injuries later that day at a nearby hospital. Soon after, his techie mates gathered to remember him. But one of them, San Francisco-based Eugenia Kuyda, wasn’t content with memories alone.
She set out to use the artificial intelligence chatbot she was developing to handle restaurant reservations to recreate her dead friend digitally. Mazurenko hadn’t been a big user of social media networks, but was a prolific texter. Kuyda used those texts, dating as far back as 2008, and others Mazurenko had sent to his circle of friends to help a neural network learn his personality.
“She fed them into an AI chatbot, made that available as an app so that anyone can download Roman Mazurenko for free on their phones and chat with a dead man,” Patrick Stokes, associate professor of philosophy at Deakin University last week told the audience at SXSW Sydney, the Asia-Pacific arm of SXSW, a media and music conference and festival.
Stokes, author of the book Digital Souls: A Philosophy of Online Death didn’t find the chatbot to be a compelling “simulacrum” of a human being.
“But I didn’t know him, and also I don’t speak Russian,” he clarified.
For Kuyda and many of Mazurenko’s friends however, chatting with the digital Mazurenko was therapeutic. The chatbot featured an option to ask him for advice and some did so, receiving brief, pragmatic responses, delivered in the distinct style of their lost friend.
Death is a binary concept, says Stokes. You’re either alive or dead. But advances in technology mean we may need to revisit our definition of death. The transhumanist movement dreams of the day when we can upload our consciousness to a computer and live on long after our physical demise.
“Well, that’s all fine in practice, but will it work in theory?” Stokes asked. Meaning that it may well be possible one day – Elon Musk’s company, Neuralink, is one of a handful in the race to develop computer-brain interfaces. But whether we should do it is another question entirely.
What rights and responsibilities would you have when you are a fully-digital entity? What right does your family and friends have to reanimate you as a digital avatar to salve the grief they feel over your absence?
Society doesn’t have answers to these questions yet. We memorialise Facebook pages to remember the dead, but the digital estate we leave is now capable of so much more. In the era of ChatGPT, iRoman could be recreated as a much more realistic digital avatar than was possible in 2015.
“It’s really important to differentiate between living on for ourselves, and living on for other people,” Stokes points out.
If you eventually have the choice to live on in a conscious state on a computer server after your body dies, at least you have some agency over your future. But we could see the data of the dead increasingly mined to bring them back, including realistic video and audio representations of them.
I’m happy for my digital remains to turn to dust. I certainly don’t want to be revived as a chatbot. But I think of someone like broadcaster Sir David Attenborough, whose documentaries have been hugely influential to me and millions of people around the world. Aged 97, we not likely to get many more films from the natural history icon – unless he lives on as the digital star of the show.
I’d be very happy with that – but I’m thinking about what I want. That decision is for Attenborough to make.
So, the next time you update your will with instructions about what you want done with your body after you die and how your assets are to be divided up, give some thought to your digital estate. Do you want to leave one at all or direct your family to press delete?
