Identical twins
Face ID’s accuracy fails when dealing with identical twins.(1) Sometimes, it cannot tell the difference between them.(2)
Apple, Face ID security guidance
VeryAI is now Very.
While both Very and Face ID aim to verify users at scale, there are some key differences with how and what they authenticate. Face ID is Apple-only and is generated to unlock specific devices. Very’s palm biometrics can be used on any operating system and are tied to you, no matter where you log in.
Very and Face ID both aim to solve similar problems: How do you confirm that you’re the same person who registered an account? But these two technologies take different approaches, each with their strengths. We chose palm biometrics because they are a stronger, more private, and scalable way to verify a user’s authenticity.
Read the litepaper
Face ID’s accuracy fails when dealing with identical twins.(1) Sometimes, it cannot tell the difference between them.(2)
Apple, Face ID security guidance
It’s not just twins. Other relatives like parents or siblings can also cause false identification and unlocking.
Apple, Face ID security guidance
Even with Face ID set up for a mask, Apple says that face coverings can reduce accuracy.
Apple, Face ID security guidance
Unlike Face ID, which is tied to specific smartphones and reliant on infrared cameras, Very is OS-agnostic and can be used by more than 85% of the global smartphone market. Face ID can be used by only up to 22% of smartphones worldwide.
A single palm verification has an up to 1 in 10 million chance that a user will be falsely identified,(3) a 10x improvement over Face ID’s 1 in 1 million chance.(1) By scanning both palms during the verification process multiple times, Very can uniquely identify every single person on Earth, with room to grow.
| Features | Very | Face ID |
|---|---|---|
| Accuracy | High1 in 10 million chance a user will be falsely identified.(3) Palms are unique, even in twins. | Medium1 in 1 million chance a user will be falsely identified. Cannot tell the difference between identical twins. |
| Scalability | HighCan be used on most modern smartphones | HighBut limited to iPhones and iPads. Face ID is tied to a device, not a user. |
| Security | HighThere are few digital datasets to pull from | MediumDeepfake faces can be created from digital footprint |
| Privacy | HighPalms are not currently linked to identifying databases | LowPhotosets featuring faces are very common |
Yes, Very can be used to verify users on both Android and iOS devices; it can be used by more than 85% of the world smartphone market. Face ID can only be used on iOS devices.
Integrating Very into your stack is simple with support from the Very team, and can take only a few lines of code, depending on your integration model. Once you receive your credentials, integration can take 10 minutes or less (integrating User Verification with OAuth using the separate Very app) up to a few weeks (a more integrated, customized solution specific to your platform’s needs using the Very SDK).
Very accounts are based on a user’s biometrics. Unlike Face ID, it is not tied to a specific device. A user can log into their Very account on any device.
Yes, Very follows the EU General Data Protection Regulation (GDPR) for all personal data we process:
• We ask for your permission before we collect your palm scan or other biometric data.
• We keep data only as long as we need it, and our Privacy Policy lists exactly how long.
• You can ask to see, correct, or delete your data at any time, and we respond within the time GDPR requires.
• We have appointed Rickert Rechtsanwaltsgesellschaft mbH in Bonn, Germany as our official GDPR representative in the EU.
• We only work with vendors that meet GDPR standards. You can see the full list at trust.very.org/subprocessors.
Business customers can also sign our Data Processing Agreement (DPA). For more details, read our Privacy Policy or email legal@very.org. We have also appointed a Data Protection Officer, reachable at dpo@very.org.
Very works with outside researchers (like the Human-AI Collaboration Lab, located at Northwestern University) to stay ahead of the latest advances.
Very introduces a number of protections against spoofed or deepfake results, including: randomized motion checks, camera-frame analysis, checks for issues like palm positioning, lighting, blurriness, and more, server-issued protections like one-time nonces to protect against replay, and others.
It depends on how you set up your integration with Very. If you use the Very SDK, you can implement Very’s palm biometrics directly within your platform. For the OAuth and ZK integrations, users must use the Very app.
Book your introductory call to start your integration.