Biometrics and other forms of pattern recognition are efficient and intelligent. It is not only possible to run image recognition searches (in real time, via all available spimes and sensor feeds) but to search for patterns such as gait, sounds, colors, emotive displays, traffic, crowd movement, etc. Kinesics and behavioral analysis even allow sensor searches for people exhibiting certain behavioral patterns such as suspicious loitering, nervousness, or agitation.
Information and files online come with hidden data about their creation, alteration, and access. A photo’s metadata, for example, will note what gear it was taken on, who took it, when, and where, as well as who accessed it online, though such metadata may be easily scrubbed or anonymized.
The combination of abundant computing, archived data, and ubiquitous public sensors enable intriguing correlations to be drawn from data that is mined and collated. In the midst of a habitat emergency such as a terrorist bombing, for example, the ID of everyone in the vicinity could be scanned, compared to data archives to separate out those who have a history of being in the vicinity at that particular time with those remaining checked against databases of criminal/suspect history and their recorded actions analyzed for unusual behavior.
Real-time translation of audio and video is available from open-source translation bots.
A significant percentage of what people do on any day or in response to certain situations conforms to routines, enabling easy behavioral prediction. Muses take advantage of this to anticipate needs and provide whatever is desired at the right moment and in the right context. The same logic applies to actions by groups of people, such as economics and social discourse, making prediction markets a big deal in the inner system.