Facial Recognition

A visual-based software that first uses a built-in camera to identify the presence of a human face, then compares the personal features of an individual to a database to successfully determine the identity of that person. Accordingly, this tech could be used by the authorities to quickly identify suspects, or as a method of authorizing or granting access to certain services.

Biometric technologies could measure individual characteristics by identifying several data points and microexpressions. By using a variety of sensors to track biomarkers, a set of algorithms would then analyze and categorize the input into recognizable human emotions. When powered by deep-learning technology, the algorithms would be able to crosscheck emotions with other data such as gender or age and provide insightful, comprehensive analytics.

This type of technology could improve the responses of devices that use contextual awareness or measure the effectiveness of ads. It could also be useful to enhance human-machine interactions in the case of home assistants, AI mentors, or general-purpose robots. It would help artificial agents in recognizing the identity of different household members and their specific moods and emotional states. The ability to read and understand human emotions is an essential feature in improving the socialization between people and machines. For doing this, a microphone is employed, and it would be possible to identify words and phrases in spoken language, thus converting speech into a machine-readable format. In an industrial context, this could make communication more fluid between humans and machines in a manufacturing plant, allowing engines to understand the human context in conversations, thereby enhancing the safety of workers.

Facial recognition is already part of the daily life of many people as it is available in many smartphones' cameras, and it is currently used to validate a user's identity, add custom-made filters, or even perform payments. Now it is also making its way into XR headsets, and paired with artificial intelligence and powerful computing power, they can translate the user's facial expressions in real-time and hyper realistically to their digital avatar, improving the social experience in games and virtual environments.

In the public sphere, there are already services that provide access to shared cloud-based networks in which every member can edit and add subjects of interest. When one of those subjects is identified in real-time by any camera grouped into the system, the owner of that camera is notified, creating a face-watch among a network of private businesses. It could be used to identify thieves, or even to find missing children using algorithms that project how a child’s face would develop in the years following their disappearance.

Of course, services like this raise quite a variety of privacy concerns. Many countries throughout the world are already implementing this technology as a means to observe and monitor their citizens' behavior and attitude in public spaces, thus providing accurate data to ensure police enforcement in cases of crime or unlawful acts. Governments and institutions are legitimized to use this technology. Nevertheless, the amount of sensitive data displayed and stored by these systems raise ethical apprehensiveness of who is accessing this information or at what point individuals could control what is being shared about them.

Future Perspectives

Being able to recognize and categorize human emotions, in conjunction with mixed reality contact lenses or augmented reality glasses, could be particularly useful in assisting neurodivergent individuals, such as those along the autism spectrum to better navigate social situations. On the other hand, as urban areas increasingly turn into surveillance states, many individuals experience an ever-growing level of discomfort. In order to dodge pervasive identification systems, some people are making use of special masks or anti-surveillance fashion accessories to confuse facial recognition software.

Sometime in the future, robots could be able to recognize emotions, not only employing surveillance systems. When artificial agents become autonomous from human interests, they would be able to read an individual’s facial expressions and analyze gestures and vocal patterns, and after that, be capable to reproduce these very emotions with the same level of complexity as their analytical ability. However, even if they are fully prepared to reproduce all emotional states, the matter of actually feeling emotions is an entirely different milestone, and robots would still be mechanical objects. In this sense, the act of mimicking human emotions on the side of robots could be considered as a valid form of deceiving and manipulation to persuade humans.

Image generated by Envisioning using Midjourney