The Password Fallacy

We've been taught to protect our digital lives with passwords. "Make it strong," "Don't reuse it," "Change it every 90 days." It's a ritual we perform religiously.

But there is a fundamental flaw in this logic: Passwords are designed to be changed. If a hacker steals your password, you reset it. The breach is contained.

Now imagine a thief steals your face. Or your fingerprint. Or your voice.

You can't reset those. You can't grow a new face. You can't change your fingerprint. Once your biometric data is stolen, it is compromised forever.

This is the reality of our Hardware Leak series. We are moving from the sensors that watch and listen to the sensors that identify. And the stakes have never been higher.


What Are Biometrics?

Biometrics are biological measurements used to identify individuals. Your device uses them to unlock itself, but they are also harvested to build a profile of who you are.

The Big Three:

  1. Facial Recognition (FaceID / 3D Facial Mapping): Your phone doesn't take a photo of your face. It projects over 30,000 infrared dots onto your skin to create a mathematical model of its depth—the curve of your nose, the distance between your eyes, the contour of your jaw. This model is stored as a hash on the device's secure enclave. In theory, it never leaves the phone. In practice, apps that request "FaceID" access are interfacing with that model, and the data pipeline extends further than most users realize.
  2. Fingerprint Scanning (TouchID / Capacitive Sensing): Your fingerprint is a unique ridge pattern. The scanner doesn't "image" your finger; it maps the capacitance differences between the ridges and valleys of your skin. This creates a high-resolution topographic map that is, again, stored as a mathematical representation.
  3. Voice Print (Speaker Recognition): Your voice has a unique frequency, pitch, cadence, and resonance pattern. Speaker recognition systems extract these features to create a "voiceprint"—a mathematical signature that identifies you. Modern generative AI models can clone your voice from as little as three seconds of audio.

The "Unresettable" Risk

Why is this data so dangerous?

1. Permanence If a database of passwords is breached, the company forces a reset. If a database of biometric templates is breached, there is no reset button. Your child's face is now a public key. If that key is stolen, it can be used to unlock accounts, bypass security, or impersonate them for the rest of their life.

2. Irrevocability You can change a password. You can change a phone number. You cannot change your DNA. You cannot change your iris pattern. You cannot change your gait. If a malicious actor obtains your child's biometric data, they possess a permanent key to their identity.

3. The Generative AI Connection Biometric data is the raw material for generative AI cloning. The more data a company has on your child's face and voice, the easier it is for bad actors to create a realistic synthetic replica. Imagine a scammer calling your child's school, sounding exactly like you, asking for an early pickup. Or a deepfake video of your child saying something they never said, circulating on social media.

This isn't hypothetical. Voice cloning services already exist commercially. Facial reanimation technology can puppet a person's face in real time. The biometric data your child's device collects today is the fuel for these systems tomorrow.


The Hidden Collection: It's Not Just for Unlocking

You might think, "I only use FaceID to unlock my phone. That's safe."

It's not that simple.

  • App Permissions: Many apps request "FaceID" or "Camera" access not just to unlock, but to "verify identity" for banking, shopping, or social features. Each request extends the data pipeline beyond the secure enclave.
  • Background Scanning: Some apps scan faces in the background to "tag" people in photos or to feed classification models that categorize users by age, gender, or emotional state for targeted advertising.
  • Third-Party Sharing: Your biometric data is often shared with third-party vendors, data brokers, and machine learning training datasets. Once it leaves your device, you lose control over who trains what on it.

The Neurodivergent Layer: When Biometrics Don't Fit

Biometric systems are designed for a neurotypical baseline. They assume the user can hold still, look directly at the camera, press a finger flat against a sensor, and speak in a clear, measured tone. For many neurodivergent children, these assumptions break down.

Sensory Overload and Face Scanning: The infrared dot projection used by FaceID is invisible to the human eye, but the act of being scanned—holding still, staring directly at the device, waiting for the "green check"—can be intensely uncomfortable for children with sensory processing differences. The demand to "hold still and look at the screen" can trigger anxiety, avoidance, or shutdown.

Motor Coordination and Fingerprint Readers: Children with dyspraxia or fine motor coordination differences may struggle to place their finger flat and still enough for TouchID to register. Repeated failed attempts create frustration and a sense of failure around a security feature that is supposed to help them.

Interoception and Voice Commands: Children with interoception differences may struggle with the regulated breathing and vocal control required for voice-based authentication. The demand to "speak clearly and naturally" into a device can feel performative and exposing.

The Practical Impact: When biometric systems fail to accommodate neurodivergent users, the fallback is almost always a passcode or password. This is actually a good thing from a privacy standpoint—passcodes are more secure than biometrics in many scenarios because they can be changed. But the experience of "failing" at biometric authentication can reinforce a child's sense that technology wasn't built for them.

What to Do: If your child struggles with biometric authentication, don't force it. A strong alphanumeric passcode is a perfectly valid—and often superior—alternative. The convenience of FaceID is not worth the cost of your child's comfort or the permanence of their biometric data in a corporate database.


The Action Plan: Biometric Lockdown

  1. Audit: Go to Settings > Face ID/Touch ID (iOS) or Settings > Security (Android). See what is enabled and which apps have access.
  2. Disable Non-Essentials: Turn off FaceID/Fingerprint for apps that don't strictly need it (e.g., games, social media, shopping).
  3. Use Passcodes: For high-security apps (banking, email), use a strong alphanumeric passcode instead of biometrics. Yes, it's less convenient. Convenience is the enemy of sovereignty.
  4. Physical Protection: Remember—if someone holds your child's phone to their face while they are asleep, they can unlock it. Biometrics are convenient, but they can be used against you. A passcode cannot be extracted from a sleeping child's finger.

The Permanent Key

Biometrics are a double-edged sword. They are convenient, but they are permanent. Once your child's biometric data is out there, it belongs to the data pipeline, not them.

In Part 3, we will bring it all together with The Hardware Lockdown Checklist—the final step to reclaiming control over every sensor, every permission, and every data point.

Call to Action: Check your biometric settings today. How many apps have access to your child's face? How many have access to their fingerprint? The answer might surprise you.


Quick Reference - Biometrics

Biometric Type What It Is The Risk Best Defense
Facial Recognition
(FaceID)
3D map of facial features (dots, depth) Permanent; cannot be reset; used for deepfakes Disable for apps; use passcode for device unlock
Fingerprint
(TouchID)
Topographic map of skin ridges Permanent; can be lifted from surfaces Disable if motor coordination is an issue; use passcode
Voice Print
(Speaker Rec)
Frequency, pitch, and cadence analysis Clonable by AI from seconds of audio Disable voice assistants; use Safe Word protocol
Iris/Gait
(Emerging)
Eye patterns and walking style Hard to detect; impossible to change Avoid devices that enable these by default