Meta patents AI that lets dead people post from the great beyond

Nothing is certain, they say, but death and taxes. But a new idea from Meta could add social media to that list.

The tech giant was granted a patent in December that would allow it to simulate a user via artificial intelligence when he or she is absent from the social network for extended periods, including, “for example, when the user takes a long break or if the user is deceased.”

The patent covers a bot that could simulate your activity across Meta’s products, including Facebook, Instagram, and Threads—making posts, leaving comments, and interacting with other users. It could even, potentially, communicate directly with people via chats or video calls, the patent reads.

Andrew Bosworth, Meta’s chief technology officer, is listed as the primary inventor, and the patent was first filed in November 2023. A Meta spokesperson tells Fast Company the company has “no plans to move forward with this example.”

Withdrawing from a social media platform can affect “the user experience of several users,” the patent reads. “The impact on the users is much more severe and permanent if that user is deceased and can never return to the social networking platform.”

Creepy? Sure seems it. Unprecedented? Not as much as you might think.

In 2021, Microsoft obtained a patent for a chatbot that would let you “talk” with dead people, both loved ones and celebrities. Like Meta, Microsoft said it had no plans to use the technology—and Tim O’Brien, Microsoft’s general manager of AI programs at the time, said in a social media post he agreed it was “disturbing.” Meanwhile, startups like Eternos and HereAfter AI let people create a “digital twin” that can engage with loved ones after they have passed away.

Meta first publicly discussed the concept of a chatbot for the dead about two-and-a-half years ago, when founder Mark Zuckerberg, in an interview with podcaster Lex Fridman (in the Metaverse, of course), said, “If someone has lost a loved one and is grieving, there may be ways in which being able to interact or relive certain memories could be helpful.” Zuckerberg did note, however, that the technology could become “unhealthy”.

Meta’s take on a post-mortem chatbot would analyze “user-specific” data, including posts, voice messages, chats, comments, and likes, to build a sense of who the person was. It would amalgamate that data into a digital persona designed to mimic the user’s activity.

The bot would identify that any responses were not actually generated by the user, the patent says, but rather were the result of a simulation.

Now, there are some hurdles Meta doesn’t mention in the patent. What people say in a direct message to a close friend or loved one isn’t necessarily meant for wider consumption. Picture, for instance, one spouse venting to the other about how frustrated they were with their child after some “terrible twos” or teenage incident—only for that child to later be told by the bot how much they annoyed their now-dead loved one.

After all, AI has yet to grasp social niceties, or when silence or a white lie is better than the truth.

Presently, when someone dies, Meta offers several options for survivors. The page can be permanently removed (assuming you have the necessary paperwork, such as a death certificate), or it can be turned into a memorial, where people can read past posts and leave messages of their own.

As unpleasant as the topic is, Meta has good reason to think about death. One study predicts that by 2050, the number of dead users on Facebook will outnumber the living. By 2100, there could be more than 4.9 billion dead profiles on the platform.

View the full article