A case that has gone viral across Chinese social media this week has forced a national conversation about the ethics of AI-generated digital twins. A family in Shandong province commissioned AI developer Zhang Zewei to create a digital replica of their son, who died in a road accident, to comfort their elderly mother. The mother — who is in her 80s and has a serious heart condition — has been having regular video calls with the AI version of her son for approximately one year, unaware that he died, according to reporting by South China Morning Post citing the original investigation by Litchi News.
The family’s motivation was compassionate: they feared that the shock of learning about her son’s death could be fatal given her cardiac condition. Zhang Zewei built the digital twin using photographs, videos, and voice recordings provided by the family, training a model capable of conducting real-time video conversations that mimic the son’s appearance, voice, and conversational style closely enough to deceive a grieving elderly woman.
How the Digital Twin Was Built — and What It Can Do
Zhang Zewei, who operates a small AI development studio, described the technical process in interviews with Chinese media. The model was trained on a dataset of the deceased son’s photos and videos spanning several years, combined with voice recordings that allowed the system to replicate his speech patterns and vocal characteristics. The resulting system can conduct live video calls, respond to questions about family history and shared memories, and adapt its conversational tone to match the emotional register of the interaction.
The level of fidelity achieved is striking. The mother, who interacts with the system regularly, has reportedly expressed that the calls bring her comfort and that she feels her son is “still there.” From a purely technical standpoint, the case demonstrates how accessible and capable AI-driven digital twin technology has become, a development that was largely theoretical five years ago is now being deployed in ordinary Chinese households by individual developers working with off-the-shelf tools.
The Consent Problem at the Heart of Grief Tech
The case raises a consent problem that existing Chinese regulations do not adequately address. The deceased son never consented to the creation of a digital replica of himself. The elderly mother is interacting with an AI system under a false premise. She believes she is speaking to her living son. Neither the deceased nor the primary user of the system gave informed consent to the arrangement.
China’s Cyberspace Administration has issued guidelines on AI-generated content and deepfakes, requiring labeling of synthetic media in most contexts. But those guidelines were designed primarily for public-facing content such as news, entertainment, social media, rather than for private, intimate applications like grief technology. The Shandong case falls into a regulatory gap that the CAC has not yet addressed, and the viral spread of the story has put pressure on regulators to act. The broader regulatory challenge connects to China’s broader efforts to regulate the country’s booming “digital human” industry.
A National Debate About Where AI Ethics Ends and Family Love Begins
The public reaction in China has been divided in ways that reflect genuine moral complexity. Many commenters expressed sympathy for the family’s predicament and admiration for the ingenuity of the solution. Others raised serious concerns about the ethics of deceiving a vulnerable person, even with benevolent intent, and about the long-term psychological effects on the mother if she eventually discovers the truth.
Legal scholars and AI ethicists quoted in Chinese media have called for specific legislation governing the creation and use of digital twins of deceased individuals, including requirements for consent from the deceased’s estate, disclosure obligations to users, and restrictions on commercial exploitation of grief technology. The Shandong case is not isolated — a small but growing industry of Chinese developers now offers digital twin services for deceased individuals, operating in a legal and ethical grey zone that regulators have not yet entered. Without consent requirements, disclosure standards, or quality controls, the market is developing without the guardrails that would normally accompany a technology with such direct and profound effects on vulnerable people.
The Shandong case will almost certainly accelerate regulatory action. The CAC has moved quickly in response to viral social controversies before — the deepfake regulations of 2022 and the generative AI rules of 2023 both followed periods of intense public debate triggered by specific high-profile cases. The grief tech industry should expect that a regulatory framework is now a matter of when, not whether, and the companies operating in this space would be well advised to begin developing their own ethical standards before Beijing imposes them.