Facebook may soon start targeting entire families with ads based primarily on the photos they post online, having just won a patent detailing such a technology. The Menlo Park, California-based social media giant drafted a concept of an online service that would identify the members of any particular household by analyzing their photographs, profile them into demographics, and feed its conclusions to Facebook’s ad-serving system. The solution utilizes computer vision and machine learning, meaning it not only has the potential to be extremely consistent but also never truly stops improving and gets better at its task with every session, successful or not.
The vanilla version of the program is meant to be trained using deep learning techniques before being tasked with analyzing family photos, though it’s been envisioned to consistently rely on similar artificial intelligence principles. Besides facial recognition, the ad-targeting mechanism could rely on photo tags and basic Facebook profile information to learn of individual user habits and relationships within any given household, the patent documentation explains. Knowing where to look is another important component of the envisioned solution, which is why Facebook wants to have it rely on a robust prediction model that would both suggest new data points to analyze and predict future behavior in order to assist the advertisement delivery network.
Background: Facebook is no stranger to trying to innovate in the world of targeted advertising seeing how almost the entirety of its business model is based on such solutions. The company may still be extremely careful with implementing technologies that target entire families in the future as it has recently been facing significant backlash over the manner in which it handles user data and approaches the topic of digital privacy. Between the Cambridge Analytica scandal from earlier this year and several smaller such debacles that followed, Facebook has now attracted unwanted attention from legislators and regulators in many parts of the world, including its home country.
Earlier this week, incoming Chairman of the House Judiciary Committee’s antitrust panel — David Cicilline — said “Facebook cannot be trusted to regulate itself,” adding that the newly elected Democratic body will be looking into the company’s recent transgressions after it’s sworn in this January. Several members of the European Parliament expressed a similar sentiment earlier this year as part of a hearing that saw Facebook co-founder and Chief Executive Officer Mark Zuckerberg face questions over the Cambridge Analytica incident. Amid all of that turmoil, the social media juggernaut still went ahead with launching its first set of smart displays that follow users around the room during video calls, describing them as privacy-oriented devices, which some industry watchers found laughable.
Impact: Facebook’s newly published patent raises yet another series of privacy concerns surrounding the social media company’s efforts to monetize its technically free service. While the timing of its publication is less than opportune given the number of scandals Facebook endured in recent months, it’s also not something the company could have easily affected given how it filed for the patent 18 months ago, before the Cambridge Analytica debacle and other such troublesome episodes. Given the current state of affairs, Facebook likely won’t be in a hurry to start mass-collecting and analyzing data for family advertising profiles, provided that ever happens.