Top 3 Ways Organizations are Leveraging Facial Recognition

Face detection/recognition has been a growing field for the last few years. With advancements in computer vision, companies have been able to leverage powerful algorithms to help solve a variety of problems. However, there have been debates about the balance between benefit to society vs the cost of people’s privacy.

Here are a few ways that organizations are using face detection.

Tracking and Monitoring

One of the most controversial areas right now concerning organizations using face detection technology is law enforcement. Some people might be worried given the plethora of futuristic movies depicting an overseeing government watching and tracking down every person’s movements.

minority report.gif

This is a complicated issue because while this will induce valid privacy concerns from the public, this can also help save lives and prevent potential crimes, or find criminals after a committed crime.

The way that law enforcement agencies are able to use these algorithms is to typically purchase services from a different company.

Image for person tracking.8a06c906b7eb728a69ee26b41933babc1067e48f.jpg

Amazon is one popular option that offers its Rekognition API service to help law enforcement agencies in identifying faces from uploaded video, social media, and other sources.

While on its own, the algorithm cannot actually do face recognition (with the exception of celebrity faces that it has already trained on) — law enforcement agencies are able to provide their own database of labeled photos which then links to Rekognition and allows it to identify people’s faces that appear in the agency’s database.


Another popular way that companies are using face detection/recognition is by adding people’s faces as a layer of security for accessing devices.

face recognition.gif

Using Apple as an example, they have completely replaced the industry standard of using a fingerprint scanner to unlock a smartphone. In 2017, Apple announced their new Face ID technology that would replace their fingerprint “Touch ID” method of unlocking the iPhone. All subsequent iPhones have used Face ID as the only option to quickly unlock the phone. This suggests how confident Apple is in their face recognition technology that they made it the sole way (aside from manually inputting a PIN code) to access the phone.

Microsoft has also made use of face recognition as part of their “Windows Hello” biometric-based system. This system allows users to unlock their Windows 10 device by either a fingerprint, iris scan, or face recognition.


Unlike the law enforcement example where organizations wanting to utilize face recognition have to look towards other companies for solutions, both Apple and Microsoft develop their methods in-house. From a privacy perspective, Apple states that all of the data collected using Face ID is encrypted and never leaves the device ¹.

Microsoft also states that face data stays on the device, and is never uploaded anywhere. The way that “Hello” works is it initially takes a reading of the face where it stores that data as a vector representation — Microsoft makes a note to state that they don’t even keep an actual image of the face anywhere. It then uses an algorithm to match the scanned face data to one that is registered to the device. Regarding accuracy, Microsoft states that false positives (stranger unlocking a device) are likely less than 1/1,000,000 of the time, while false negatives (failing to unlock the first time) can occur in less than 5% of cases.

Both of these uses of face recognition are useful in everyday life, and seem to impact us minimally from a privacy perspective, especially since the companies involved are trying their best to keep our data from being shared online.

Customer Experience

While the previous examples focused on security and safety, some companies are opting to use face detection for fun instead.


In 2015, the immensely popular photo sharing app, Snapchat, released a “Lenses” feature which uses face detection to augment one’s appearance. Whether it is to apply some skin smoothing, wear a funny hat, or adding dog ears, people have had a lot of fun using this to change their look. Over the years, Snapchat has improved the feature by adding impressive tracking, as well as an amusing way to swap faces with another person, or even swapping faces with a saved photo on your device.

On the semi-controversial side, there has been a viral app called FaceApp going around which has an impressive algorithm which can show people how they would look like when they are much older, among other features. Due to the popularity of the app, people started asking questions about how the photos uploaded to the app are being used. Rumors also went around about the app potentially uploading all the photos stored on the phone and using them as training data for their algorithms. This, however, has not been backed up by evidence and could potentially have been started by public panic3.

Industries implementing face detection/recognition will have a variety of different problems they will encounter — a major problem is how to maintain privacy while still providing value to the public. Consumers should be able to hold organizations to a standard of keeping data secure.