Can we humans live beyond our biological lifespans? In light of rapid advances in artificial intelligence research, it’s conceivable that it will soon become possible. To some extent, it already is: to create a new type of legacy in the form of a virtual entity, such as a chatbot or an avatar, that acts as a digital replica of someone. This is the question Mi3 assistant professor Jurgis Karpus grapples with in a recent paper he co-authored. Here, Jurgis explains more about whether it’s possible to say that the original person continues to exist in their digital replica.
Science fiction novelists and filmmakers have been exploring the idea of creating digital replicas of persons for quite some time. Would it be possible to upload our brains and minds into computers and to thereby extend our existence beyond our physical bodies and normal biological time? Could advanced chatbots, when trained on large amounts of data about our psychology—our desires, beliefs, fears, and worldviews—accurately represent our views (and us) to future generations?
Some have taken up this idea seriously and are developing technologies to do just that. But how should we see such digital replicas of human persons? Should such replicas be considered mere counterfeits of people, or is there something more to them? Could they become genuine extensions of their replicated persons? In a recently published paper, myself and fellow philosopher Anna Strasser argue that in some sense they can.
We have developed an argument building on the work of Derek Parfit, a British philosopher who previously presented a prominent philosophical understanding of what personal identity and personhood—the stuff that makes us us—is. According to this view, our personhoods are not “all-or-nothing” things or states of being. Our personhoods are rather graded—sometimes more of whatever makes us us persists through time and other times less. As a result, utterances such as “I am now a different person from the one I was ten years ago” should not be thought of as mere metaphors—they can be stating facts.
Building on this idea, we argue that if large amounts of data about our psychology—about the stuff that makes us us—are used to create digital representations of us using tools powered by artificial intelligence, then bits (or, in our words, slivers) of our personhoods are indeed genuinely extended to such entities. They become part of us. While the idea might be exciting, scary, and to some (perhaps many) outright outrageous, we say we ought not worry. While some of our personhoods may indeed be extended to such digital entities—digital replicas of us—those bits or slivers of personhood would be quite “slim” and would on most occasions not constitute what we ordinarily consider to be a full person. Also, no one would personally think it enticing to continue their existence as a digital entity without consciously experiencing that existence—something a machine could not do.
Nevertheless, certain ethical and legal implications are important. AI-powered tools have been used to create “new” songs by Nirvana and Amy Winehouse, were consulted to complete Beethoven’s unfinished 10th symphony, and taught to compose chorales in the style of Bach. Would these late artists, had they been alive today, have grounds to claim ownership of such AI-enabled creations? Should we care about whether these artists would approve of the new music created in their name? We argue that the answer to both questions might well be “yes.” The late artists, or rather their living representatives, could now bring a new argument into a courtroom and demand their cases to be heard.
Check out the original study and Jurgis Karpus’ profile and LinkedIn.
References:
Karpus, J., Strasser, A. Persons and their Digital Replicas. Philos. Technol. 38, 25 (2025). https://doi.org/10.1007/s13347-025-00854-z