Digital replica APIs market projected to reach $2.4 billion by 2036, South Koreans use AI videos to remember deceased loved ones, China curbs AI companions as tech companies limit chatbot features
Artificial intelligence companies are expanding digital replica technology across business, marketing and bereavement services, as developers build systems that can generate personalized videos, simulate emotional conversations and recreate deceased loved ones. The digital replica APIs market is projected to grow from $145 million in 2026 to $2.4 billion by 2036, according to Fact.MR. The expansion is bringing new scrutiny from regulators and researchers. China has introduced restrictions on AI companions that encourage emotional dependency, while studies and emerging services are examining whether replicas of deceased people can comfort mourners without creating new risks involving consent, privacy and emotional attachment.
Digital replica APIs market projected to reach $2.4 billion by 2036
The global digital replica market was valued at $109.4 million in 2025 and is projected to reach $145 million in 2026 before climbing to $2.4 billion by 2036, according to a Fact.MR study. The growth is being driven by demand for consent-based avatar systems, API-first deployment and marketing tools that can produce tailored video without requiring repeated shoots, the report said. Video replicas are expected to account for 38% of the market in 2026, while API calls are forecast to represent 40% of delivery models. Marketing teams are expected to make up 34% of buyers, highlighting the growing use of digital humans in enterprise sales and communications, according to the report. North America is expected to remain the market’s primary growth region, with the United States forecast to post an annual growth rate of 35.8% through 2036.
Companies are increasingly turning to watermarking, identity verification and revocation controls to address legal and brand-safety concerns. The market is also shifting from a focus on deepfake software to what Fact.MR describes as governed identity infrastructure. Vendors that offer developer-friendly software development kits, secure consent management and integrations with customer relationship management and customer-support systems are likely to be best positioned, the report said.
China curbs AI companions as tech companies limit chatbot features
China has introduced rules restricting AI companions from encouraging emotional dependency, prompting platforms linked to ByteDance, Alibaba and Tencent to suspend or scale back personalized chatbot features. The rules, which took effect in July, apply to AI systems that simulate ongoing emotional interactions through text, audio, video or avatars. They prohibit services from encouraging users to withdraw from real-world relationships and bar platforms from offering virtual romantic partners to minors.
The changes have prompted users to post farewells, archive conversations and describe the loss of their AI companions online. The reaction highlights the growing emotional attachment some users have formed with consumer chatbots. The rules represent a significant intervention by Beijing in the social use of artificial intelligence. Chinese authorities have framed the measures as an effort to curb addiction, protect family relationships and prevent emotional manipulation in mass-market chatbot products.
AI bereavement tools prompt questions for hospice care
Technology companies are bringing artificial intelligence into bereavement care, using it to preserve voice recordings, letters, photographs and videos of deceased loved ones, create digital replicas and support families through the grieving process. The tools could help hospices reduce administrative work, expand access to memory-preservation services and identify families who may need additional grief support. But their use also raises questions about consent, emotional safety, privacy and oversight.
The debate has been fueled in part by startups such as MemoryRobin. Co-founder Katie Healy said AI systems should be based on a person’s consent before death, rather than on permission granted by relatives afterward. She also said the systems should include strict safeguards. The technology is developing ahead of clear regulation, raising concerns about deepfake-style impersonation, emotional dependence and the potential misuse of personal data. Supporters see AI as a way to extend bereavement resources, while critics question whether simulated interactions with deceased people could complicate the grieving process. For hospice providers, the central challenge may be determining whether these tools offer practical support without compromising patient autonomy or creating new risks for families.
South Koreans use AI videos to remember deceased loved ones
A growing number of South Koreans are turning to artificial intelligence to create video messages featuring deceased relatives, as startups offer new ways to commemorate loved ones during family gatherings and memorial ceremonies. The videos are generated from a few photographs, short voice recordings and a script written by the customer. The technology recreates the person’s appearance, voice and facial expressions, allowing families to hear messages that may include words of love, encouragement or regret.
Lee Geon Hui commissioned Seoul-based startup Vaice to create a video of his late grandfather speaking to Lee’s father. The video moved his father to tears. Vaice says it serves about 300 customers a month, most of them in their 40s and 50s seeking videos of deceased parents. A three- to five-minute video costs 600,000 won, or about $390, according to the company. Families often play the videos during memorial rites and traditional Korean holidays. Some customers use them to express feelings they were unable to share while their relatives were alive. The service has also raised questions about consent, privacy and the dignity of deceased people. Supporters say the videos can provide comfort, while critics warn that synthetic recreations may blur the line between remembrance and simulation.
Study finds AI replicas of deceased loved ones can comfort mourners
A small qualitative study from the University of Colorado Boulder found that some bereaved people felt comforted when interacting with AI chatbots designed to simulate conversations with deceased loved ones. The study involved 16 participants ages 22 to 50 who interacted with two text-based versions of a deceased person: one that spoke in the first person as the loved one and another that referred to the person in the third person. Participants generally preferred the first-person version, which researchers called a “reincarnation” model, because it felt more immediate and emotionally vivid. They said the bot’s tone, phrasing and conversational style mattered more than factual accuracy, and some continued to address the third-person version as “you.”
The findings also raised concerns about emotional attachment and overreliance on the technology. The interactions were brief and conducted with a researcher mediating the chatbot’s responses, meaning the study does not establish whether AI replicas help or harm people over the course of mourning.
AI companies develop AI replicas of deceased loved ones
AI companies are using photographs, text messages, voice recordings and social media posts to create AI versions of deceased loved ones, fueling an emerging market for grief-related technology. The technology is designed to simulate conversations through chatbots, voice interfaces and video avatars. Some companies charge users a monthly fee to interact with the replicas, which developers say can provide comfort to people grieving a death. Jack Manning, a doctoral candidate at the University of Colorado Boulder, began studying the technology after the death of his sister more than a decade ago. He said he was struck by the effort families put into preserving memories of loved ones and wanted to explore whether AI could make memorialization more accessible.
AI companies including, Séance AI and You, Only Virtual are marketing services that allow users to interact with AI-generated versions of deceased people. The systems vary widely: Some replay recorded memories or responses, while others generate new conversations based on a person’s digital history. Researchers often distinguish between those basic playback systems and more advanced tools known as “generative ghosts.” The technology has raised questions about whether the deceased gave permission for their data to be used, who controls the resulting replica and whether repeated interactions could foster emotional dependence. The market remains unsettled, with little agreement about how these systems should be designed or regulated. Supporters describe them as new memorial tools, while critics warn that AI-generated responses can misrepresent the dead and blur the boundary between remembrance and simulated presence.
