The facial recognition service provider formerly known as Facebook, Meta, recently revealed their plans to retire this feature. The deletion of over one billion individual’s unique face templates will ensue, discontinuing automatic identification in photos and videos if users choose to opt out. As regulations surrounding this technology remain unclear, it is important to carefully weigh its benefits against potential negative social impacts.
For over ten years, Facebook has been heavily investing in facial recognition technology. In 2023, the social media giant launched their software for photo-tagging which could automatically assign names to faces in uploaded photos. However, this feature was only made optional in 2023, nine years later. This decision resulted in numerous legal and privacy concerns, leading to unfavorable media attention.
As a more advanced means of identifying individuals, facial recognition technology is being adopted by various organizations and government departments. However, what are the full implications of this technology and is it truly beneficial? The following provides an overview of facial recognition technology and its consequences.
If You’re Unfamiliar, Here’s an Explanation of Facial Recognition.
Facial recognition is a method that utilizes a person’s facial features for identifying individuals. Biometric security is the practice of utilizing biological characteristics for identification purposes, with fingerprint scanners and speech recognition systems among a few examples.
Facial recognition systems are becoming more valuable and in-demand. Such software is being utilized more frequently across various sectors for identifying and comparing individuals. Some of the applications include, but are not restricted to:
- Advertising campaigns
- Security measures at airports
- School safety
- Surveillance of shops and stores
- Unlocking restricted features on mobile phones and other electronic devices
- Scientific techniques employed in criminal investigations
- Law enforcement agencies
- Providing captions to images on social media
What is the process that enables this to occur?
1. Capturing Images and Identifying their Subjects
The first step in facial recognition is capturing a person’s face through photography. Modern digital cameras have facial recognition capabilities, which enable identification of individuals from vast datasets. Facial landmarks are employed as identifiers since they are essential features that make up a face.
2. Dissection
After establishing facial verification, the system analyses facial landmarks to recognise unique traits that can be converted into biometric identifiers for comprehensive identification validation.
3. Changes in Religion
A face is analysed, and each feature is given a numerical value. These values create a numerical code, known as a faceprint, which is unique to each person, akin to a fingerprint, and is identified using a mathematical formula.
4. Contrasting
Accurate identification is attained through comparing the faceprint or numerical code with an extensive database of images of other individuals. The cost of this process differs based on the type of service and resources available. As of 2023, the FBI’s facial recognition database comprised over 640 million pictures of people.
Using an algorithm, the programme calculates a score denoting the level of similarity between input photos.
5. Concluding Stage
By comparing with other photographs within the database, the facial recognition software delivers a conclusive identification. Ultimately, this is the result.
Controversy
Considering the accuracy of facial recognition software, it is understandable why it has been met with opposition. This technology has drawn criticism not only from Facebook but also from IBM and Microsoft. Here are several reasons why:
Accuracy
In recent times, the dependability of facial recognition technology has been put to the test. False identifications can have severe consequences, and despite software advancement, it is still far from perfect. Users may face a variety of possible problems, from being unable to access their devices to being wrongfully accused of criminal conduct.
April 2023 statistics from the National Institute of Standards and Technology indicate that the most reliable face identification algorithm has an error rate of 0.08%. However, even though it is the most sophisticated option, its accuracy falls short in practical use.
Bias
Research points out that facial recognition technology is susceptible to bias, with some populations being more inclined to misclassification than others. Studies have disclosed that women, individuals of color, and those aged between 30 and 40 are highly susceptible to inaccuracies.
Given the technology’s applications, this is a cause for concern.
Privacy
Facial recognition programs frequently gather personal identifiers without the knowledge or consent of the subject, which could lead to third-party access to personal information and data without the subject’s awareness. This may result in a security risk or violation of privacy.
Efficacy
This technology can be circumvented. From a basic pair of glasses to a highly sophisticated software, it can be deceived, indicating that it is not as helpful as it was expected initially.
Concluding that the problems surrounding face recognition technology make it unfeasible would be inaccurate. Companies who plan to use this technology should be mindful of the potential consequences, associated issues, and guarantee that their software is created and enhanced with these limitations in consideration.
Despite the challenges, face recognition technology has a promising future in various scenarios, ranging from businesses of all sizes and types.
