printer friendly version

AI-driven identity verification for access control

The human need to protect assets has always existed. Whether it be classified information or personal property, limiting access to these resources based on identity has existed since modern civilisation started.

Throughout history, we have seen examples of the lengths people will go to conceal their valuables from the wrong people. In 100 BC, Julius Caesar was known to transmit messages to generals using codes and cyphers, an early form of encryption. Fast forward to 1960, when Fernando Corbató first used passwords to secure computer files.

We have evolved, though, and in 2022, identity verification is not conducted by people or people-driven applications, but by artificial intelligence (AI).

As the machine learns, complex algorithms compile data and then use the data to make better decisions. You have likely already interacted with an AI-driven application today if you have conducted a Google search, had a song recommended by Spotify, or simply unlocked your iPhone.

How does AI-driven identity verification elevate security?

Identity verification powered by AI can be applied in different ways, perhaps the most popular being through facial authentication. Facial authentication is an actively recognised process where a person’s face is their identification verifier. We will expand upon this concept later, but from here on out, we will discuss AI-driven facial verification as it relates to access control and security.

Multi-factor authentication

Multi-factor authentication refers to an extra layer of authentication required for a user to gain access. Multi-factor authentication aims to increase the difficulty for an unauthorised user to gain entry into a secured facility. AI facial verification is the ideal choice as the first layer of authentication due to its ease of use and touchless requirement. To create a multi-factor process, facial authentication can then be paired with additional access control credentials like keycards, pin codes, or other biometric factors.

Data security

A sound access control system that employs facial authentication technology will ensure that the data collected by the system is protected. Any information processed by the technology should be encrypted and used solely for authentication purposes. When it comes to users’ faces and verification, enrolled users should be given the option to opt-in for the verification, allowing individuals a level of control over their profiles.

Frictionless solution

Facial authentication solutions combine advanced AI and 3D sensing technologies with ease of use to create a frictionless, touchless experience. The deployment of this technology in an access control system keeps users and administration moving. Users do not have to stop and take out an access keycard; the admin does not have to worry about manually enrolling users and issuing cards.

Artificial intelligence learns as it goes, so user profiles are built based on a few device interactions, making enrolment a breeze.

Zero Trust environment

Derived from the basic security principle of ‘never trust, always verify’, Zero Trust means designing security around the assumption that your company’s associates, applications, and devices cannot be trusted. Access control systems with facial authentication automatically employ this principle since the camera detects intent to approach the door in an attempt to authenticate each face. Again, the automated processes functioning behind the scenes determine if the user is to be trusted and, therefore, granted access or not.

Authentication vs recognition: Why does it matter?

Facial authentication and facial recognition are often thought of as being synonymous. While similar, the two terms serve different purposes and employ different processes (see box) on page 52. It is essential to understand these variations, as distinct ethical and legal concerns are associated with facial recognition.

For example, the deployment of facial recognition technology rightfully gives some cause for concern. Companies that employed this technology have come under fire from many who believe privacy rights were violated during the process. Why? Because facial recognition technology collects biometric data from users who are often unaware that said data was collected. This data is then stored in databases shared with or sold to third parties for use in surveillance or targeted marketing campaigns. To protect the privacy of citizens, legislation, such as the Illinois Biometric Information Privacy Act (BIPA) and California Consumer Privacy Act (CCPA), was created to limit the use of facial recognition in these types of applications.

Compare this process with facial authentication, where a user actively engaging with the system provides the same or similar biometric data. In this process, the user is cooperative and aware they are using the biometric system. These systems provide tools that allow companies to administer the biometric profiles to adhere to local legislation and protect the privacy of the enrolled individuals.

The non-cooperative use of user data associated with facial recognition is often what people think of when they hear ‘facial authentication’. Hence, making the distinction between the two is essential. Users should be confident that their data is protected and that their identity is not used for profit or otherwise.

Why is anti-tailgating important?

In the world of access control, tailgating occurs when an individual, either with or without authorisation, follows an authorised user through a secured door. To many, the simple act of holding the door open for someone is seen as harmless, polite, and considerate, but there are serious risks that come with tailgating.

In the best-case scenario, a user with access holds the door for another approved access holder, but this creates a bad habit, and there is often no way of knowing if a tailgater has been given access or not. In a worst-case scenario, individuals with malicious intentions are granted access to a secured area, negating the access control system altogether.

Luckily, modern advances in access control have allowed organisations to mitigate these risks by deploying anti-tailgating technology.

Anti-tailgating technology, like the Alcatraz AI Rock solution, can prevent and detect tailgating. Automating this process via an intelligent access control system can save companies millions compared to the cost of guards and installing physical barriers to deter tailgaters.

Devices equipped with anti-tailgating technology detect tailgating by identifying individuals in real time as they approach an entrance. The device will attempt to authenticate all faces seen within the device’s field of view to determine if each person is authorised. Suppose an unauthorised user follows an authorised user through a door. In that case, the solution will identify that second user as a tailgater and issue a tailgating alert to the admin of the access control system. If applicable, the alert will also include a photo of the unauthorised person, allowing security or management to identify the offender and take immediate action.

To prevent tailgating, the system can also be configured to send an alert to the access control system in the form of a unique credential that can only give access to authorised users. This is especially useful for facilities that have multiple entry points.

Administrators can then leverage the data constantly gathered by the access control system to identify tailgating hotspots, or the employees most likely to allow tailgating. Armed with this knowledge, organisations can improve workplace operations and reinforce good physical security habits.

Benefits of video at the door

The application of video at the door systems is not new. Think of popularised home security solutions like Nest and Ring video doorbells. In addition to providing homeowners with a real-time view of who is at their doorstep, the videos captured by the camera have been used to catch thieves and criminals in the act. Even the mere presence of a video camera at eye level is often enough to deter potential trespassers with nefarious intentions.

These video-at-the-door solutions are great for homeowners, but organisations need something a little more robust. This is where the intersection of video surveillance, access control, and facial authentication comes into play. When these capabilities are rolled into one system, the system owner and users all benefit.

Artificial intelligence and facial authentication

Alcatraz AI leverages artificial intelligence and facial authentication to provide physical security solutions for sensitive areas.

For too long, physical security has forced organisations to choose between traditional technologies that keep environments safe but sacrifice movement, and modernised options that keep workers moving but limit security.

Our vision at Alcatraz is to eliminate this trade-off through a solution that delivers on both fronts; cost-effective for organisations, actionable for managers, and intuitive for users. We do this by employing artificial intelligence, facial authentication, and real-time analytics.

The Rock, Alcatraz’s flagship product, is an award-winning access control solution designed to provide frictionless, highly secure entry into physical locations. The Alcatraz Rock seamlessly integrates with existing access control systems to offer multi-factor authentication. Some features of The Rock are listed below:

• Multi-sensor technology.

• PoE powered.

• Wiegand/OSDP outputs and inputs.

• Tamper detection.

• Thoughtful user interface.

• Enrolment intelligence.

• Tailgating intelligence.

• Mask enforcement.

Facial authentication vs recognition

Facial authentication

Definition: The biometric verification of a consenting user’s identity.

In layman’s terms: A user consents to the use of their unique biological characteristics to confirm they are who they say they are.

Example: Unlocking your iPhone or Android device with your face.

Facial recognition

Definition: The gathering of biometric data from non-consenting individuals for comparison with ‘like-data’ in a database.

In layman’s terms: Users’ faces are scanned, often unknowingly, with the intent of matching a face to a similar one in an existing database of faces.

Example: Casinos scanning faces of entrants looking for known, blocked cheaters.

This is a shortened version of a white paper from Alcatraz.ai, represented in South Africa by C3 Shared Services. Download the full version at https://alcatraz.ai/resources/ai-driven-identity-verification or use the short link www.securitysa.com/*c3ss1

Share this article:

Further reading: