Meta patents AI to keep dead users posting

Imagine scrolling through your feed and seeing a new post from a friend who passed away months ago, complete with their signature phrasing and humor.

That Black Mirror-esque scenario is officially a patented concept at Meta. As reported by Business Insider, the tech giant was recently granted a patent for an AI system designed to simulate a user’s social media activity, including keeping their accounts alive and interacting after they die.

A digital ghost in your DMs

Originally filed in 2023 and granted in late December, the patent details how a large language model can recreate a person's online presence.

The system works by training on user-specific historical data. That covers pretty much everything you do on Facebook or Instagram: past updates, comments, private chats, voice messages, likes, and typical behavioral patterns.

According to the filing, the AI could step in when a user is "absent from the social networking system." While that could technically mean taking a social media detox, the patent explicitly notes that the impact of absence is "much more severe and permanent" when an account holder has died.

Once active, the reconstructed digital persona could publish new posts, reply to friends, and send direct messages. The document even sketches out advanced territory, referencing models capable of mimicking the person’s voice and likeness for audio and video calls.

For now, Meta is downplaying the discovery. A company spokesperson insisted that Meta has "no plans to move forward with this example," emphasizing that patents are frequently registered simply to lock down intellectual property that may never become real features.

The Zuckerberg connection

While Meta claims it is just a paper concept, CEO Mark Zuckerberg has openly toyed with the idea before.

In a 2023 interview with Lex Fridman, Zuckerberg suggested that AI could eventually help people interact with memories of loved ones who have died. He noted that Meta would likely have the capacity to build AI replicas of individuals, though he stressed that user consent is mandatory and that such decisions "should ultimately be your call."

Meta is hardly the only player fascinated by grief tech. Several startups have already launched apps allowing families to create digital avatars of deceased relatives. Every rollout inevitably triggers intense debates over digital consent and the psychological toll of conversational ghost bots.

Why it matters

We think this patent pulls back the curtain on the logical, if unsettling, extreme of engagement-driven platforms.

From a user perspective, this is not something anyone should be rushing to adopt. Interacting with an algorithmic recreation of a lost friend feels less like closure and more like an uncanny valley trap. Without airtight guardrails, you open the door to horrific identity disputes, unauthorized account takeovers, and emotional manipulation.

For Meta, the technical rationale is obvious. Real people log off, lose interest, and eventually pass away, taking their network value with them. A simulated digital persona, however, stays active forever, keeping comment sections moving and social graphs intact.

Filing defensive patents is standard Silicon Valley procedure. But actually deploying a bot that chats from beyond the grave would be an unprecedented ethical minefield. Until platforms can guarantee rock-solid consent mechanisms and genuine human empathy, some digital doors are much better left closed.