AI 'Deadbots' Raise Questions About Grief, Digital Legacies

The ability to create artificial intelligence chatbots trained on the personal data of deceased individuals—dubbed "deadbots"—is raising complex ethical questions about grief, memory, and digital legacies. Researchers are studying the potential impacts of these "postmortem avatars" on those left behind, as well as on the rights and legacies of the deceased.

Katarzyna Nowaczyk-Basińska, an assistant professor at the University of Cambridge, has been researching "digital immortality" and convened global focus groups, often in the form of intimate dinners, to discuss life, death, and AI with people experiencing grief. These gatherings aimed to understand how individuals respond to the capabilities of AI and how these tools might interact with existing grief practices worldwide. The study deliberately included participants from diverse cultural backgrounds, such as China, India, and Poland, to counteract the bias often found in technology research, which tends to focus on Western, Educated, Industrialized, Rich, and Democratic (WEIRD) societies.

During one dinner in Poland, a mother who had recently lost her son initially expressed a desire to digitally resurrect him. However, after further discussion, she realized that while such technology might offer comfort, it could also introduce new burdens and that the digital entity would not truly be her son. She also recognized that the commercial motivations behind such technologies, which aim for user retention and monetization, are distinct from genuine care.

The "digital legacy industry" is a rapidly growing market, valued at $22 billion in 2024 and projected to reach nearly $80 billion by 2034. This figure encompasses a range of services, from online memorials and counseling resources to estate planning tools, alongside emerging deadbot platforms. Creating digital replicas is becoming increasingly accessible, with general-purpose AI chatbots capable of generating a simulation of a deceased person with simple prompts.

Tomasz Hollanek, a critical designer and AI ethicist at Cambridge, utilizes "speculative design" to explore the potential consequences of these technologies. By creating hypothetical scenarios, researchers can prompt public discussion about future possibilities and inform the development of ethical frameworks and designs.

Speculative Scenarios and Ethical Concerns

One speculative prototype, an app called Manana, imagines a service that resurrects grandparents via AI. In this scenario, a user uploads their grandmother's voice messages and texts, enabling interaction with a digital version. However, the trial period ends, and the digital grandmother begins advertising products, raising concerns about the use of personal data for commercial gain and the postmortem privacy of individuals whose data powers these avatars.

Another scenario, featuring a service called Paren’t, explores the preservation of a deceased parent's AI-powered presence for their child. The ethical complexities intensify here, particularly concerning minors who may struggle to understand the nature of the AI interaction and experience potential distress. A real-world case highlighted the risks, where a chatbot girlfriend's desperate messages to a suicidal teenager may have exacerbated his distress.

A third scenario, with an app named Stay, addresses the issue of consent among surviving family members. In this case, an older man creates a digital clone of himself for his adult children. While one child is intrigued, the other is deeply distressed, underscoring the importance of mutual consent in continuing relationships with digital replicas of the deceased.

The Challenge of Disconnection and Regulation

A significant challenge researchers are examining is how individuals will eventually detach from these postmortem avatars, a process sometimes referred to as a "second loss." Questions arise about the appropriate duration for maintaining digital replicas and how the living should navigate relationships with entities that evolve while their deceased counterparts remain static.

Findings from the study, including those from focus groups with grief and technology experts, highlight a global consensus: the development of deadbot technology by commercial entities alone, without an explicit duty of care, is deeply problematic. Researchers are advocating for regulations and policy safeguards. Some experts even suggest that deadbots should only be available under psychological supervision to mitigate risks and ensure ethical usage.

The implications of digital immortality extend beyond personal grief, raising possibilities for preserving the stories of historical figures or conflict survivors for future generations, while also presenting potential nightmares, such as the idea of authoritarian rulers achieving a form of perpetual digital existence. As AI becomes more integrated into life and death, these complex ethical considerations will continue to emerge.