In the past year, with a certainly impressive amount of research and experimentation, facial and image recognition technologies have gained considerable momentum. Among the largest innovators are search engine giant Google, and the colossal online community Facebook, who have harnessed the power of revolutionary facial recognition and image recognition technologies.
Just over a year ago, Facebook released DeepFace, intending to “close the gap to human-level performance in face verification”. By merely looking at two photos, the facial recognition technology deciphers whether the photos contain the same face, irrespective of lighting and positioning of the face. With 97.3% accuracy, only second to the 97.5% accuracy of the human brain, DeepFace is an impressive piece of artificial intelligence that certainly meets its intended goal.
Similarly, Google’s image recognition technology has advanced to the point of being able to mimic human levels of understanding. Although it relies heavily on image descriptions in order to tag them, the search engine is able to identify entire scenes in photos and videos. For example, Google’s image recognition can detect a group of children playing on a jungle gym. By working with Stanford University and merging two types of neural networks, Google has made the cataloguing and searching of billions of images and videos more efficient.
While such technologies may sound fun and useful, both have stirred up considerable controversy. Some Facebook users have protested DeepFace’s preemptive and automatic tagging feature, positing that it violates their human rights. Google has had to meet concerns of some of its users, now that it’s able to track all searches and identify all new content uploaded daily. The search engine recently included a solution to delete and/or download any information users may not want tracked, whether from their search history of Gmail account.
Concerns may start to arise around Facebook’s mobile app, which enables users to take and post photos straight from it. However, if the app’s face verification automatically starts tagging the users in these photos as soon as they’re uploaded, users may feel trepidatious about posting them in the first place. Furthermore, Google has an extensive suite of mobile apps and owns the Android ecosystem, all of which could be easily integrated with image recognition. Google could integrate everything from Google Camera to Google Drive with image recognition to instantly tag photos on your mobile phone and in Google Cloud services. Overall, both companies should take stock in how users will view their apps if privacy concerns remain relevant.