Meta Patents AI That Could Keep Users Posting After Death

Did Meta patent an AI that can post after someone dies?

Yes. In late December 2025, Meta was granted a patent describing an AI system that could simulate a user’s social media activity, including posts, replies, and interactions, even after the user is deceased or inactive. However, Meta says it has no plans to develop or release this feature. The patent was originally filed in 2023 and lists Meta CTO Andrew Bosworth as its primary author.

What exactly would this AI system do?

The patented system would train a large language model (LLM) on a person’s historical activity, posts, comments, likes, tone, and interaction patterns, to simulate how they behave online.

According to the filing, the AI could:

• Reply to comments
• Send or respond to direct messages
• Like posts
• Generate new posts
• Simulate video or audio calls
• Maintain engagement during long absences

The system is designed to activate when a user is “absent from the social networking system,” including after death.

How would Meta’s posthumous AI work?

The AI would analyse years of a user’s digital history to replicate their communication style and behavioural patterns.

Step-by-step concept:

1. Data Collection:
2. Posts, comments, reactions, messages, and interaction habits.
3. Model Training:
4. A large language model learns tone, vocabulary, humour, preferences, and social patterns.
5. Simulation Phase:
6. The AI continues interacting as if it were the original account holder.

Unlike memorial pages, this system would not just archive content. It would actively generate new responses.

Why did Meta file this patent?

Meta described the system as a “continuity tool,” especially for influencers or creators who may want engagement maintained during breaks.

The patent suggests two primary use cases:

ScenarioPurpose

Extended absence: Maintain account activity

Post-death, prevent sudden platform inactivity

The filing states that the impact is “more severe and permanent if that user is deceased and can never return.”

Is Meta launching this feature?

No. Meta has publicly stated it does not plan to move forward with the AI posthumous simulation concept.

A spokesperson confirmed the company currently has no development roadmap tied to the patent.

However, patents often reflect technical experimentation rather than confirmed product plans.

Has another company tried something similar?

Yes. Microsoft patented a chatbot in 2021 designed to simulate deceased individuals, but later scrapped the idea, calling it “disturbing.”

Meanwhile, startups have explored AI “deadbots,” including:

• Replika AI
• 2wai

These tools attempt to mimic deceased individuals through AI-generated conversations.

What are the ethical concerns?

The patent raises major questions around consent, digital identity, and post-mortem data rights.

Key concerns include:

• Who grants permission for posthumous simulation?
• Would families control the account?
• Would users know they are interacting with AI?
• Could the AI generate statements the person never actually made?
• What happens to long-term data boundaries?

Estate planning experts now advise individuals to include AI usage instructions in digital wills.

How is this different from a memorial profile?

A memorial profile preserves content. Meta’s patented AI would generate new content and interactions.

Memorial Page: AI Simulation

Static archive: Active participation

Reflects past life: Continues conversations

No new content: AI-generated responses

This shifts the concept from remembrance to digital replication.

Why does this patent matter?

Even though Meta says it won’t launch the feature, the patent signals that AI is moving toward identity simulation, not just content generation.

It shows:

• AI can now replicate communication patterns
• Social platforms are exploring digital continuity
• Generative AI may extend into post-mortem identity

The technical capability exists. Whether it becomes mainstream remains uncertain.

Meta’s patented AI system proposes a future where social media accounts could remain active after death, powered by large language models trained on personal data. While the company says it has no plans to implement it, the filing highlights a growing industry interest in digital legacies and the complex ethical territory that comes with them.