Summary: Generative artificial intelligence is quietly reshaping how people experience grief, remembrance, and mortality. Digital reconstructions of those who have died can offer solace and preserve memories, but they also risk blurring the natural line between presence and absence and complicating the process of emotional closure.
Researchers and clinicians caution that depending on AI for emotional support or decision-making may weaken our tolerance for uncertainty and impede the development of empathy. Over time, these shifts could change how societies understand death, dignity, and the practice of letting go.
Key Facts
- Digital afterlife: AI chatbots and avatars can imitate deceased individuals, extending memory while complicating emotional closure.
- Psychological risk: Maladaptive or prolonged use of grief-focused AI may interfere with acceptance of loss and the reality of impermanence.
- Human connection: Face-to-face empathy, shared rituals, and community support remain central to healthy mourning and understanding of death.
Source: Kyoto University
Shisei Tei says he is awkward with technology and does not own a smartphone, yet he spends a great deal of time thinking about generative AI.
Tei approaches AI with cautious optimism. As a researcher he applies AI tools to analyze psychiatric data, and in his personal life he uses them to plan tailored hikes. At the same time, he is concerned about how AI may alter cultural attitudes toward death, a topic he explores in a chapter of the book SecondDeath: Experiences of Death Across Technologies.
“Today, I often see how AI reframes grief and remembrance,” Tei says. He acknowledges that mental health chatbots can lower barriers to care, but he warns that chatbots designed to reconstruct deceased individuals may distort how people experience death and existence.
“AI-induced virtual continuations of the deceased can comfort the living and extend memory to some extent,” Tei notes, “but they can also blur presence and absence, potentially hindering our capacity to accept impermanence.”
Tei situates his concerns within a broader intellectual history. Many cultures and philosophical traditions have long separated mind and body, a view that can support beliefs in an enduring or immortal mind. That perspective has found new expression in efforts to preserve or simulate human minds through digital means.
Based in Japan and originally from Taiwan, Tei works at Kyoto University and focuses his scholarship on bridging psychiatry, religious philosophy, and neurophenomenology — an approach influenced by biologist Francisco Varela. In his book chapter he examines death through the idea of “selfless selves,” a term linked to Varela and to thought influenced by Tibetan Buddhism.
Selfless selves describe living systems that maintain themselves through the interdependence of their parts, like cells within a body. “Selfless selves refers to being both altruistic and autonomous — maintaining one’s individuality while remaining in harmony with others and the wider world,” Tei explains.
From this perspective, people are distinct yet continuously co-creating a shared life; the self is fluid and shaped through social interaction to meet biological and social needs. Tei observes that modern AI agents display similar traits: they present various artificial identities, lack a fixed or intrinsic selfhood, and operate within networks of anonymity and interconnection online.
Yet there is a tension between this relational view of self and AI’s tendency to deliver quick, definite answers. Traditional philosophies and contemporary clinical practices often emphasize the value of accepting uncertainty, ambiguity, and loss. By contrast, AI can encourage outsourcing of decision-making and emotional labor to systems built around efficiency and optimization.
“Outsourcing decision-making or emotional support to machines risks weakening the very wisdom we aim to cultivate,” Tei warns. Human empathy, especially the kind formed through face-to-face and nonverbal communication, plays a unique role in creating belonging and conveying what it feels like to be alive. Meanwhile, solitude and sorrow can, paradoxically, nurture hope and personal growth.
Perceptions of death arise from these living interactions. For many, dying evokes a sense of connection to something larger than the individual: although a person’s biological life ends, aspects of them may continue through relationships, shared memories, and community practices. Tei argues that integrating these relational ideas into end-of-life care and communal conversations can help preserve dignity and enable acceptance of mortality.
“Death becomes certain once life begins,” Tei writes in the chapter, “and denying its anticipation risks denying life itself.” His view calls for balancing technological innovation with human-centered care, cultural ritual, and practices that cultivate tolerance for life’s uncertainties.
Key Questions Answered:
A: AI can simulate the presence of the deceased through chatbots and digital avatars, which may provide comfort but also risk blurring the boundary between living presence and absence.
A: By offering ongoing digital “continuations” of the dead, AI can disrupt the natural process of accepting impermanence and the end of biological life.
A: Outsourcing emotional regulation and meaning-making to machines could erode human empathy, diminish tolerance for uncertainty, and weaken emotional resilience over time.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The referenced book chapter and journal material were reviewed in full by editorial staff.
- Additional context and framing were added by the editorial team to situate the research for general readers.
About this AI, grief, and death research news
Author: Whitney Hubbell
Source: Kyoto University
Contact: Whitney Hubbell – Kyoto University
Image: The image is credited to Neuroscience News
Original Research: The chapter titled “Death in the Cybernetic Era: AI, Virtual Agents, and Selfless Selves” appeared on 16 October 2025 in SecondDeath: Experiences of Death Across Technologies.