Anyone who tries to edit their date of birth on Instagram from under the age of 18 to 18 or over is asked to verify their age. They can either upload their ID (which is powered by Meta) or use Yoti’s facial age estimation technology. This was first introduced in the US and has since expanded to the UK, Europe, Brazil, India, Mexico, Canada, South Korea, Australia, and Japan.
Meta’s approach to age verification
In 2019, Instagram began asking people to provide their age when signing up for the platform. Since then, they’ve continued to ask people for their birthday and have made this a requirement to get a full picture of how old people are on Instagram. Knowing how old everyone is allows them to provide different experiences to different age groups, specifically teens. When they know if someone is a teen (13-17), they can default them into age-appropriate privacy settings, prevent unwanted contact from adults they don’t know, and limit the options advertisers have to reach them with ads.
Many people, including young people, may misrepresent how old they are online, whether by mistake, because they don’t want to share their age, or because they want to access services not meant for them. Some companies have turned to government issued identity documents or credit card checks to verify age but many people, particularly teens, don’t have access to these forms of ID. The addition of Yoti’s facial age estimation technology to the existing range of solutions offered by Meta means users will be provided with more equitable options to verify age that protect peoples’ privacy.
A privacy-preserving way to verify age
Meta is using Yoti’s facial age estimation technology as one option to verify age. Yoti’s technology has been trained to estimate age by looking at facial features in an image. It was built to give everyone a secure and private way of proving how old they are. It doesn’t require any personal details or documents, and it has no way of linking a name or an identity to an image.
Yoti’s facial age estimation is built in accordance with the ‘privacy by design’ principle in the UK GDPR. No individual can be identified by the model and it encourages data minimisation because it only needs a facial image. The image is only used for the purpose of estimating age. Once a result is given, the image is instantly deleted.
How Yoti’s facial age estimation works on Instagram
If someone chooses to verify their age with Yoti, they will see instructions on screen to take a video selfie. After they take the selfie, Meta shares only the image with Yoti. No other details are shared.
The image is analysed by AI that has been trained to estimate age. To the technology, the image is simply a pattern of pixels, and the pixels are numbers. Our facial age estimation technology has been trained to spot patterns in numbers, so it learns ‘this pattern is what 16 year olds usually look like’.
The estimated age is then shared with Meta. After this, both Meta and Yoti delete the image. Yoti’s facial age estimation technology can’t recognise identity – just age, so it does all of this without ever learning who anyone is. Yoti does not use these images to train the AI further.
Yoti’s facial age estimation technology has performed over 570 million million checks worldwide, and is being used by a range of businesses and industries around the world including social media, gaming and age-restricted e-commerce.