AI Deathbots: The Rise of Digital Resurrection & Grieving Concerns

A recent Rod Stewart concert in Charlotte, North Carolina, offered a glimpse into a burgeoning, ethically fraught phenomenon: the digital resurrection of the deceased. Among his surprise guests were AI-generated images of late music legends, including his friend Ozzy Osbourne, alongside Michael Jackson, Tina Turner, and Bob Marley. The spectral reunion divided fans, with some decrying the tribute as disrespectful while others found beauty in the digital homage.

This incident is not isolated. Around the same time, a former CNN White House correspondent, Jim Acosta, conducted an interview with a digital recreation of Joaquin Oliver, a 17-year-old tragically killed in a 2018 Florida high school shooting. Created by his parents, the avatar offered them the profound solace of hearing their son’s voice once more. Similarly, Reddit co-founder Alexis Ohanian shared an animation of his late mother hugging him as a child, generated from a photograph. The deeply personal experience left him emotionally raw, highlighting the powerful, often unexpected, impact these digital echoes can have.

These examples underscore the rapid emergence of “digital resurrection,” a practice involving the creation of images, videos, and interactive bots of deceased individuals using their photographs, voice messages, and other digital footprints. Companies offering “griefbots” or “deathbots” are proliferating, prompting a surge of questions concerning exploitation, privacy, and the profound implications for the grieving process.

The technological leap enabling this trend is largely attributed to the widespread availability and user-friendliness of large language models (LLMs) like ChatGPT. As cyberpsychologist Elaine Kasket notes, these LLMs facilitate the creation of highly plausible and realistic digital representations. If sufficient digital “remains”—texts, emails, voice notes, images—exist, it becomes remarkably straightforward to construct something that feels deeply recognizable to the bereaved. What was once a futuristic concept of “virtual immortality” is now accessible and relatively inexpensive, with demand poised for significant growth. A 2023 YouGov poll, commissioned by the Christian think tank Theos, revealed that 14% of respondents would find comfort in interacting with a digital version of a deceased loved one, with younger generations showing greater openness to the idea.

The human desire to maintain connections with the departed is, of course, ancient. Throughout history, people have sought solace in cherished personal items, pored over photographs, replayed voice messages, or even engaged in seances. As Michael Cholbi, a professor of philosophy at the University of Edinburgh and author of Grief: A Philosophical Guide, observes, “Human beings have been trying to relate to the dead ever since there were humans. We have created monuments and memorials, preserved locks of hair, reread letters. Now the question is: does AI have anything to add?”

While deathbots can serve the same purpose as traditional memorials by fostering a sense of connection, they also pose potential disruptions to the grieving process. Louise Richardson, from York University’s philosophy department, cautions that ongoing interaction with a deathbot might impede the crucial work of recognizing and accommodating loss. Bereaved individuals often ponder what a loved one might have said or done in a particular situation; now, technology appears to offer a direct answer. However, this interaction risks providing a “sanitized, rosy” representation, as Cholbi points out. A creator might omit less appealing facets of a personality, presenting an idealized digital persona.

Furthermore, there is a significant risk of fostering dependency. Nathan Mladin, author of AI and the Afterlife, describes “digital necromancy” as a deceptive experience, where interaction with a machine can be mistaken for genuine connection, potentially hindering acceptance and healing. The commercialization of grief also raises concerns, with the market for digital avatars of the deceased booming, particularly in the Far East. In China, creating a basic digital avatar can cost as little as 20 yuan (£2.20), though more advanced, interactive versions can run into thousands of pounds. Estimates placed the market value at 12 billion yuan (£1.2 billion) in 2022, projected to quadruple by 2025. This rapid growth spotlights the potential for exploitation, echoing historical issues within the funeral industry.

Beyond commercial exploitation, privacy and the rights to digital remains present complex challenges. A deceased person cannot consent to their digital material being used, nor do they have a right of reply or control over their posthumous digital existence. Concerns also mount over the fraudulent use of digital material to create convincing avatars for financial gain. Some individuals are already taking proactive steps, stipulating in their wills that their digital assets should not be used in this manner after their death.

While the ethical debates surrounding “deathbots” are intense, digital avatars are not exclusively for the deceased. The wildly successful Abba Voyage show, featuring digital versions of the band performing in their prime, demonstrates the enthralling potential of such technology for living artists, generating about £1.6 million weekly. More soberingly, the UK’s National Holocaust Centre and Museum launched a project in 2016 to capture the voices and images of Holocaust survivors, creating interactive avatars to answer questions about their experiences for future generations.

Despite the current “AI hype,” Cholbi expresses skepticism about the long-term sustainability of relationships with the dead through this technology. While some may find it therapeutically beneficial or use it periodically for special occasions, he believes most people eventually reconcile with the finality of death. Mladin posits that the interest in digital resurrection may reflect a societal shift, where fading traditional religious beliefs redirect deeper longings for transcendence and everlasting love towards technological solutions. This trend, he suggests, is symptomatic of a peak modernity culture that believes technology can conquer death itself.

However, Kasket cautions against the “pathologizing” of grief by services selling these digital phenomena. While acknowledging that some find them genuinely helpful, she worries about a societal narrative that suggests an inability to cope with grief without technological intervention. Grief and loss, she stresses, are fundamental, normal aspects of the human experience, not pathologies for technology to solve.