Facial recognition technology converts facial features into mathematical representations called face embeddings, which are compared against stored data for authentication. Modern systems use deep neural networks trained on millions of images to improve accuracy. The key distinction between security levels lies in 2D camera-only systems versus 3D hardware-based systems like Apple's Face ID, which projects over 30,000 infrared dots to create depth maps and resist spoofing attempts. While facial recognition offers convenience for phone unlocking, bank verification, and airport boarding (such as India's DigiYatra), it carries significant privacy risks since biometric data cannot be changed like passwords. Security experts recommend systems that keep biometric templates encrypted and stored locally, and users should question who stores their facial data, how long it's retained, and whether they can opt out.
Deep Dive
Prerequisite Knowledge
- No data available.
Where to go next
- No data available.
Deep Dive
How They May Be Stealing Your Face! Facial Recognition Explained | Tech Today
Added:There was a time when passwords meant numbers and letters. They exist today as well, but today all of this just takes a glance. Your face unlocks your phone like that. It verifies your bank account and at some airports also, it even helps you board a flight. This face is my digital identity and it's the same for all of you right now. But here's the big question.
How does a phone actually recognize your face? Why is Apple's Face ID considered different from the face lock found on many other Android devices? And as more companies being starting to use facial recognition, how much control do we really have over our own faces? Today on the Tech Today show, we are decoding the tech that's quietly changing how we approve who we really are, where it all is used and how safe it actually is to share your facial data in today's world of artificial intelligence.
Imagine meeting someone for the first time. You don't memorize every single detail. Instead, your brain notices key features.
The distance between the eyes, the shape of the nose, the jawline, and the cheekbones.
Modern facial recognition works in a surprisingly similar way. Instead of storing a photograph, most modern systems convert your face into a complex mathematical representation, often called face embedding, which captures distinguishing facial features.
When you unlock your phone, the system compares a newly captured face embedding with the one stored during enrollment to determine whether they match.
Today's systems use deep neural networks trained on millions of facial images to improve recognition accuracy.
>> All right, get this. Not every device recognizes your face the same way. Some devices record a 2D image, while others actually build a three-dimensional map of your face. And that makes a huge difference. Something that you might not even get to know.
>> When Apple introduced Face ID in 2017, it redefined biometric recognition.
Face ID is possible through dedicated hardware called the TrueDepth camera.
The system projects over 30,000 invisible infrared dots onto your face, creating a detailed depth map.
Combined with an infrared camera and machine learning, Face ID can recognize you even in the dark and is designed to resist spoofing attempts using ordinary photographs.
Apple says the facial data is encrypted and stored only inside the device's secure enclave and the mathematical representation never leaves the iPhone or Face ID authentication.
It's one of the reasons many facial apps also use Face ID for authentication.
It's that secure. However, many Android phones, particularly budget and older models, use a simpler approach.
They rely mainly on the front-facing RGB camera to compare a two-dimensional image of your face.
While this is convenient, camera-only systems generally provide a lower level of security because they lack true depth information.
That said, not all Android phones are the same.
Devices like Google's Pixel 8 and 9 series also use infrared-assisted hardware for secure face authentication, while some premium phones from other manufacturers incorporate additional sensors.
The key distinction isn't iPhone versus Android. It's 3D hardware based authentication versus camera only face unlock.
>> All right, so today facial recognition is just everywhere. Banks use it for facial recognition during digital KYC to verify that the person who's opening the account actually matches their official identity documents. Many offices today use it for attendance systems and in India one of the biggest examples is DigiYatra.
>> The platform allows passengers at participating airports to use facial recognition instead of repeatedly showing boarding passes and identity documents.
According to the DigiYatra Foundation and the Ministry of Civil Aviation, participation is voluntary and the system is designed so that passengers biometric credentials are stored on their own device with temporary verification data deleted after the journey.
Privacy groups, however, have called for greater transparency, stronger oversight, and continued scrutiny of how biometric systems are governed.
The debate, therefore, isn't simply about the technology itself, but about trust, consent, and accountability.
>> All right, so there is absolutely no doubt that the tech is impressive and it makes life much easier and simpler for all of us. However, you also need to understand the risks of something like this. One of the biggest concerns isn't unlocking your own phone. It's what happens when companies collect millions and millions of these face images.
Probably could be for data training.
>> Perhaps the best known example is Clearview AI, a US-based company that built a facial recognition database by scraping billions [music] of publicly available images from websites and social media without asking most individuals for permission.
The company has faced a lawsuits, regulatory action, and fines in several countries over its practices.
Another concern is security. Unlike a password, you can change a PIN if it's compromised.
But, you can't change your face.
That's why cybersecurity experts recommend systems that keep biometric templates encrypted and stored locally whenever possible.
>> So, should you really trust facial recognition technology? How safe is the tech on your phone? Well, the answer to this isn't a simple yes or a no. Facial recognition isn't inherently unsafe. In fact, when implemented properly with dedicated hardware, encrypted on-device storage, and strong anti-spoofing protections, it can be both secure and convenient at the same time. The bigger question is whether the organization asking for your face data is transparent about what it collects, why it collects it, and how long it keeps it, and whether you can opt out of the system.
>> Before agreeing to a face scan, it is worth asking who stores my biometric data? Is it kept on my device or on a server? How long will it be retained? And can I delete it?
Those answers matter far more than the technology alone.
>> I've been visiting airports for such a long time, and I've been absolutely shocked knowing that many people are unaware that DigiYatra as a service is not compulsory, and it's not the government that stores your face data.
It's a private firm, and that's why you need to be aware of what's happening with your data. And yes, I understand it is right now convenient. Today it unlocks your phones, opens airport gates, verifies our identity, and is increasingly becoming the key to our digital lives. And facial recognition isn't just going away anytime soon. But the bigger challenge is making sure we know who gets access to our face data, how they are using it, and under what rules because there need to be guardrails. Because just like you won't hand your ID card to a random stranger, the same should be the case for your face data.
Related Videos

Expanding Stikbot thumbnails
leopoldshorts
2K views•2023-09-24

Digital Discrimination: Cognitive Bias in Machine Learning
redmonktechevents2974
4K views•2019-12-18

Evolutionary Approach to Clustering by Ujjwal Maulik
ICTStalks
279 views•2019-06-26

Rose Yu "Learning from Large-Scale Spatiotemporal Data"
networkscienceinstitute
2K views•2019-03-04

Stanford Seminar - Generalization through Task Representations with Foundation Models
stanfordonline
4K views•2025-07-14

Satellite-Based Wheat Yield Forecasting using GEE & Transformer Neural Network
gisrsinstitute
634 views•2025-06-15

Paradigm Shifts in Data Processing for the Generative AI Era: Robert Nishihara of Anyscale & Ray.io
GradientFlow
2K views•2025-01-02

How to Build Your Own GenAI-Based Knowledge Management System
2150GmbH
360 views•2025-06-03
Trending

Playstation NO DISC/NO BUY Fight Is Over...
DavidJaffeGames
4K views•2026-07-23

Americans Confused in Australia for 17 Minutes Straight
IWrocker
17K views•2026-07-23

Bitcoin Social Interest: Dozens of us Left
benjaminjcowen
12K views•2026-07-23

Tesla Profits Plunge & SpaceX Stock Continues Fall
TheJohnJohnstonLounge
6K views•2026-07-23