AI tech could let you speak to your loved ones after their death - how it works
As surreal as it might sound, speaking to relatives beyond the grave could become a reality.
A new artificial intelligence tool that would allow people to “take part” in conversations with a person after they have passed away is in development, with Microsoft issuing a patent for the technology.
How does it work?
Using tech that sounds straight out of an episode of Black Mirror, the chatbot would potentially allow users to converse and interact with a past or present entity. First reported by the tech news site, Protocol, the tool wouldn’t just allow you to chat with a late family member, but also an acquaintance, celebrity or a historical figure.
The chatbot would pick up the voice and information of a relative by using videos, voice recordings, any letters, and social media posts. It could potentially then be used and interacted with via a smartphone, or other smart device, like a Google Home.
It is currently unknown what Microsoft plans to do with the technology, or whether it plans to release a chat bot after the issuing of the patent.
Other similar technology
There are currently a number of start up companies in the US working on similar ideas to the Microsoft plans.
Eterni.me is a California firm that creates animated avatars of people using data from social media profiles. This is similar to the hologram of Kim Kardashian’s late father, which was recently given to her as a birthday present by her husband, Kanye West.
More similar to the Microsoft chatbot is HereAfter AI. Its creators are currently developing an app which uses voice recordings to talk and reminisce in the persona of a dead person.
Users of the app would be able to speak to a former loved one and ask a question like: “Mum, tell me about my first steps.” They would then hear their mother’s voice telling the story.
The waiting list of people to be interviewed for inclusion in the app is already several hundred long, and they will be asked questions about their lives, with responses put into different sections.