Google’s ‘audience attendance’ patent analyzes people and their body language, as they consume content, to see how interested they are. If you are a little concerned for your privacy, I don’t blame you. Although the invention here differentiates between identified individuals that A) have agreed to participate in a media study, and B) have not agreed to participate, the patented claims (the invention that government gave them patent protection over) appear to treat both types of viewers similarly. In other words, regardless of whether or not you have agreed to participate, the claimed process analyzes users’ body positions to estimate user interest, and transmits related data to a server system. This data can be used for numerous purposes, for example, to automatically provide recommendations to users.
The patent states that “In accordance with some implementations, the client system (FIG. 10, 1002) estimates user interest in presented media content analyzing captured visual data to determine the body position of the one or more users of the client system (FIG. 10, 1002) (1306). For example, the client system (FIG. 10, 1002) determines that a user is facing away from the display screen. The client system (FIG. 10, 1002) estimates that the user has a low level of interest in the presented content and records the estimation.”
To Google’s credit, there are some implementations that are described where they discard ‘private’ information. “The visual data is captured in the vicinity of the client device and thus may include images and information the user wishes to keep private. Thus, this visual data is never transmitted off the client system (FIG. 10 1002) and is removed from the system after some set period of time. For example, the client system (FIG. 10, 1002) removes visual data from memory 4 hours after the information is first captured.”
Thus, although Netflix and other streaming platforms can presumably gather data about what people have selected to stream, this data lacks insight as to whether the people are actually paying attention to the streamed content. In this case, a camera mounted on a TV or mobile device can tell whether a user is paying attention to the streamed content, and estimate low interest if the user’s body language suggests such, even though the content is streaming for the user.