Facial Recognition Technology: A Boon or Bane
The rise in public safety concerns over the years has accelerated the use of touchless access controls in most applications. The pandemic further accelerated this established dynamic to be integrated into all industry verticals making Facial recognition software the most preferred biometric benchmark. From Google’s image tagging feature to Apple iPhone’s FaceID unlock feature, Facial recognition is finding its foothold in most business sectors due to the innumerable options this multifaceted discipline has to offer.
Facial recognition technology (FRT) has revolutionized customer business interactions, by analyzing their behavior and catering to their preferences and interests. For instance, frequent customers are identified, and their prior purchase experiences are analyzed, to promote deals and discounts based on age, gender, and other related information. Customer emotions are recorded, and their reactions to new products, before releasing them, are predicted. Businesses implementing FRT provide personalized experiences with minimal or no physical intervention enforcing user-friendly experiences.
Facial recognition is one among many forms of biometric identification. Other types include fingerprints, iris scans, voice recognition, digitization of veins in the palm, and behavioral measurements. The need for frictionless access controls and easy accessibility has popularized the adoption of FRT in the last two decades. According to a recent global survey, the facial recognition market is projected to grow to a value of approximately US $8.5 billion by 2026.
How it works
FRT uses AI and biometrics to map facial features from an image or video and compares the information with a database of known faces to find a match. This identification is used to access an application, system, or service and it works like a face scanner or face analyzer. The mapping is done for identification, verification, categorization, or live detection. The process involved in FRT mainly comprises the following steps:
1. Face capture and the preprocessing phase where the detected face from the source image or video is aligned to process and analyze further.
2. Face extraction phase where facial features like the distance between the nose and mouth, the distance between forehead to chin, the shape of lips and ears, etc., that are unique to a person, are extracted to form a faceprint, unique to every individual.
3. Face Match or Recognition involves matching the faceprint with a repository of images to find a match. The results can be subsequently evaluated for accuracy.
From providing an additional layer of security to surveillance systems to authenticating access to buildings, online services, payment portals, etc., the possibilities of FRT applications are endless.
Unlike other identification solutions such as passwords, and verification by email, Biometric facial recognition uses unique mathematical and dynamic patterns that make this system one of the safest and most effective ones. Thanks to artificial intelligence (AI) and machine learning technologies, face recognition systems can operate with the highest safety and reliability standards in real time.
Applications of FRT
From the phone in your pocket to the security cameras you can spot in a mall, FRT is used everywhere for different purposes. Individuals and organizations use FRT to unlock personal devices, access private records, enter venues without showing tickets, and even aid blind and low-vision individuals. Let us look at some other applications of FRT:
Law enforcement agencies: FRT can be used with CCTV to identify missing children, suspicious behavior, and criminals.
Healthcare: FRT in healthcare assists in patient identification, record access, early threat detection, and diagnosis.
Business: Employees' office entry access and their management can be done through FRT—embedded access control and surveillance systems respectively.
Marketing campaigns: FRT can be used to identify and analyze regular customers and ideate targeted marketing campaigns for specific audiences.
Retail: FRT can be used to scan shoppers’ faces, personalize their experience and identify potential shoplifters.
Airports: For faster check-in and digital onboarding purposes. Also to identify potential threats.
Schools: The use of FRT in schools and colleges will help in monitoring students and identifying intruders on campus.
Social media: For image tagging.
Advantages of FRT
FRT can decrease the need for human interaction and thus increase efficiency. Owing to its multiple applications, FRT for Identification or authentication has several benefits:
High Security: As faceprints are unique, implementing FRT leads to secure access controls and avoids risks. This is especially useful for remote entry access.
Easy Integration: Implementing FRT in already existing Surveillance systems is easier as the infrastructure is already available and also reduces operational costs.
Faster Processing: With advancements in Deep learning and AI algorithms, FRT applications process fast and lead to accurate results.
User experience: FRT offers a personalized and smooth user experience, avoiding the need for time-consuming wait times.
Compliance: FRT is considered the only acceptable standard for remote identity verification for high-risk operations like opening bank accounts or signing contracts, etc.
Challenges
Despite the innumerable applications and benefits that come along with implementing FRT, it is important to understand the limits of facial recognition AI. There are some obstacles that hinder the full-fledged adoption of FRT in business and consumers are treading with caution while using FRT-enabled systems.
In 2015, Google Photos mistook an African-American person for a gorilla in a relatively controlled laboratory setting. In the real world where the settings cannot always be controlled due to movement, masks, crowded spaces, and other unavoidable factors, it is tough to vouch for the reliability and accuracy of Facial Recognition Systems. As the world of AI and ML is getting bigger, deeper, and more advanced leading to path-breaking discoveries, certain technical, legal, and ethical complications are developing rapidly. Privacy concerns of consumers are just the tip of the iceberg.
- Technical challenges like bad camera quality and image resolution, poor lighting, and positioning can lead to low levels of accuracy.
2. There are possibilities of FRT being ineffective when people age and the captured image does not match with the database images. Similarly, wearing a mask, sunglasses, or even certain makeup can make facial recognition less accurate.
3. Bias might creep in due to race, color, gender, demographics, and many other factors making the dataset and algorithms inaccurate to provide desirable results. Bias mainly affects people of women and color.
In July 2020, the National Institute of Standards and Technology (NIST) conducted independent assessments to confirm these results. It reported that facial recognition technologies for 189 algorithms showed racial bias toward women of color.
4. Anyone who can access the data can violate privacy.
For instance, stalkers could perform reverse image searches on pictures to gather personal information about an individual, and scam people.
5. Facial images can be exploited for identity theft.
For instance, researchers were able to hack into Apple’s face ID within 120 seconds at the annual Black Hat hacker convention in Las Vegas.
6. Potential ownership issues might surface as people tend to give up the right to ownership over images of their faces when agreeing to social media privacy policies. Most social media websites have set privacy policies where the user needs to agree to their use of personal information if need be, before opening an account.
The use of FRT has faced severe repulsion globally. Several states in the US have either banned this technology or denied any governmental, particularly police usage.
Social media giant Facebook settled a $650 million class-action lawsuit in Illinois over collecting photos not publicly available for facial recognition.
In India, the lack of law governing such technologies’ use has created a public outcry, with civil society organizations demanding a ban on using FRT. There is a constant fear of manipulation of information and an increase in the possibilities of data errors leading to mistaken identity or false accusations of theft or fraud. Personally Identifiable Information (PII) is valuable and prone to misuse by unauthorized parties. Consumers trust companies to keep their data confidential.
The ongoing debate of convenience over privacy, paired with the grappling competition in the market is forcing businesses to heighten the security features and capabilities of these systems. Companies need to strike a balance between innovation and privacy and abide by the General Data Protection Regulations (GDPR) to ensure customer awareness about the safety of their personal information. A few measures to enhance security and accuracy in FRT are listed below:
· Good camera quality, accurate enough to work with different image variations.
· Advanced enough to detect people through masks and other occlusions.
· Ability to analyze 3D imaging.
· Difficult to exploit by criminals.
· Incorporate Fair and Unbiased data.
Conclusion
As the internet is getting people closer and communicating faster, the expansion of FRT has become a prominent global issue. It is important to understand the impact of FRT applications to analyze their effects and societal implications. Technology is not always an unbiased tool and a tool for good. While FRT has many potential benefits, it also brings significant privacy concerns. In a world where faces are relied upon to confirm identity, whether in person or on a video call, understanding what is real is critical to security and privacy.
The application of facial recognition increasingly allows companies to achieve higher accuracy, and subsequently improve business outcomes. With effective legal, security, and privacy measures in place, businesses can relinquish control to customers and envisage a seamless, congenial, and secure user experience for them.
References
https://www.computer.org/csdl/magazine/co/2023/01/10008928/1JIoGc4mLE4
https://metacept.com/challenges-in-implementation-of-facial-recognition-technology/
https://learn.g2.com/ethics-of-facial-recognition
https://internetfreedom.in/problems-with-facial-recognition-systems-operating-in-a-legal-vacuum/
https://us.norton.com/blog/iot/how-facial-recognition-software-works
