Meta Patents an AI Model That Can Imitate User Activity During Absence

What the Patent Is About

At the end of December, Meta received a patent describing a large language model capable of imitating human activity on social media when a user is absent. The patent was originally filed in 2023, and its author is listed as Andrew Bosworth, Meta’s Chief Technology Officer.

According to the document, the AI model can simulate user behavior if a person takes a long break from a platform or even in cases where the user has passed away. The core idea is to preserve continuity of interaction for followers and contacts.

How the Technology Works

The model would be trained on user-generated data to understand behavioral patterns: what content a person likes, how they comment, and how they communicate in private messages. Based on this, the AI could like posts, leave comments, reply to messages, and even simulate audio or video calls.

This approach aligns with current market trends: according to Statista, over 4.9 billion people globally use social media in 2024, and more than 60% of content engagement depends on consistent activity. A prolonged absence often leads to sharp drops in reach—sometimes up to 40% for creators after a 30-day inactivity period.

Potential Use Cases and Concerns

Meta suggests this could be useful for influencers and creators whose income depends on engagement but who also need breaks. At the same time, a company spokesperson stated that Meta does not currently plan to move forward with this concept, emphasizing that patents do not necessarily lead to implementation.

Still, the idea raises ethical and philosophical concerns. Studies show that over 70% of users feel uncomfortable with AI-generated interactions impersonating real people, especially in sensitive contexts such as posthumous activity.

Ethical and Legal Implications

Legal scholars warn that such technologies go far beyond technical innovation. Professor Edina Harbinja from the University of Birmingham notes that this affects not only law, but also deep social and moral frameworks. Questions around consent, digital identity, and post-mortem data rights remain largely unresolved.