Facial recognition has become a commonplace feature on smartphone and cameras, but recognition software far surpasses these capabilities. While computers have been “learning” since the birth of Pattern Recognition in the 70’s, advancements in sorting algorithms have increased the interest in Artificial Intelligence (AI) over the past decade. Essentially, AIs contain a manufactured neural network, or convolutional neural net, that are fed massive amounts of information in order to create their own classification system. That is to say, after an AI has been shown enough photos of human faces, for example, it can find human faces by itself. 

Google has used this process to create what they call “Google Goggles,” which is a way to search the internet by taking a photo. For example, if you were visiting Portland, OR for the first time and took a photo of one of the city’s many bridges, the Google Goggles would be able to identify that landmark. However, the product only works when the landmark, famous artwork or celebrity face is already in the program’s memory; otherwise, the best such software could do would be to say that the object you photographed was a bridge. Similarly, Facebook is able to more accurately recognize and tag users in photographs if they have a number of pictures already uploaded to the site. 

Amazon is currently working on their own software, named Rekognition, which similarly recognizes faces, various emotions and thousands of inanimate objects. Amazon has said that product may be used to enhance security on mobile devices and apps, and also learn to self-sort data in order to make “smart marketing,” targeted at a specific demographic. Self-teaching software is at the core of AI, which has grown by leaps and bounds over the past several years, including Amazon’s Alexa, as well as Lex — an offshoot that is used to build conversational chat bots. 

As more and more users begin utilizing such AIs, their information stores grow and they continuously refine categories, becoming more accurate, or “smarter.” The process of sorting and recognition can do much more than deliver results about photographs, however. Google has plans to use AIs to take in vast amounts of information about companies and employees, in order to pair individuals and job openings.