Meta filed a patent for a system designed to continuously record a user's voice and surrounding environment. The U.S. Patent Office published the patent on July 2, 2026.

The patent describes an apparatus that monitors a user and their surroundings to generate workout recommendations. It uses an emotional-state machine learning model to interpret verbal and nonverbal cues from transcribed audible communications.

According to the filing, the system tracks when a user laughs, sighs, their physical location, the objects around them, and when they take medication. "In some examples, audible communications may include speech (e.g., voice data), sighs, laughter, or other nonverbal sounds associated with an expression(s), an emotion(s), or ideas," the patent said. The tone of a user's voice during communications may also be included as an input.

The patent states that an AI assistant may listen to users at predefined times to detect various types of communication. "The AI assistant may use these inputs to quantify the user's emotional state or generate other insights about the user," the document said. It noted that the system can combine audio inputs with other data to summarize emotional trends, such as "a happier emotional state associated with a particular time of day or at a time when medication is taken." The system also supports its emotional interpretations with citations to specific audio moments.

The filing claims the system increases the precision and reliability of emotional inference by aligning multimodal sensor inputs on synchronized timelines. The patent states that personal trainers cannot provide the level of precision in guidance, such as correcting a pose and body movement, that the device aims to provide.

In 2012, Meta conducted a study into emotional contagion using Facebook's newsfeed. That study involved altering the feeds of 700,000 users to observe changes in mood, and users were not informed they were part of the experiment.