Everything you need to know about our facial age estimation technology
We’re really proud of our age estimation technology, but it’s important to us that we tackle some common misconceptions. So we’ve answered some of the most important questions we get asked about age estimation. What is age estimation? Our facial age estimation technology accurately estimates a person’s age based on a selfie. We built it to give everyone a secure way to prove their age without sharing their name or ID document. This privacy-friendly approach doesn’t require any personal details or ID documents, and all images are instantly deleted once someone receives their estimated age – nothing is ever viewed by a human. It can’t link a name to a face or identify anyone. This is the difference between facial analysis and facial recognition. How does it estimate age? The technology has been trained to estimate age by looking at facial features in an image. 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’. Is it facial recognition? No. The technology uses facial analysis to estimate a person’s age without identifying or recognising any individual. When estimating age, it doesn’t cross-check people against a big database of faces, it simply estimates the age of the image presented to it. The technology never knows or learns the name or identity of a person. This has been acknowledged by multiple regulators and has prompted the ICO to update their definition of biometrics and agree that “Yoti’s age estimation tool will not result in the processing of special category data”. Where did the data come from to train age estimation? Most of it comes from Yoti app users. We use the photo from an ID document and the date and year of birth. We ask them to allow us to use the data to train our technology during the account creation process. Those users can withdraw their approval to use their data for training purposes at any time through the Yoti app. All our data for under 13 year olds has been collected specifically for the purposes of training the age estimation model and the parents or guardians have freely given consent during a specific data collection exercise. We have ensured we have obtained a balance of gender identities, images and different skin tones to minimise bias in the algorithm – we have not used any images scraped from the internet. Is the technology more accurate than humans at estimating age? It can be difficult for humans to be sure whether someone is over 18 just by looking at them. When guessing the age of another person, we tend to underestimate the age of older people, and overestimate the age of younger people. Our ability to estimate accurately tends to decrease as we ourselves get older. That’s why policies such as ‘Challenge 25’ exist, which ask customers to prove their age if they appear to be under 25. This builds in a safety buffer of 7 years given the legal age for buying alcohol in the UK is 18. Age estimation can be configured to work with legal age thresholds in a similar way. A retailer using age estimation at self-checkouts can build in a buffer of 7 years, meaning anyone estimated to be under the age of 25 will not pass and will need to prove their age another way. If we look at the accuracy of the technology when estimating the age of an 18 year old, there is a mean absolute error (MAE) of 1.22. This means an 18 year old could be estimated to be 17 or 19 years – too high or too low. This is why businesses using age estimation can use a threshold, like Challenge 25, to have greater confidence that someone under the age of 18 would not be able to buy an age restricted item. Another way to measure the accuracy of age estimation is using a True Positive Rate (TPR). This is the probability that an actual positive will test positive, meaning an 18 year old is correctly estimated to be under 23. The TPR for 13-17 year olds correctly estimated as under 23 is 99.65%. This gives regulators a very high level of confidence that nobody underage will be able to access age-restricted content, and the technology can be used in a variety of settings to strengthen age checks – retailers selling age restricted items, adult websites or content providers, and gambling terminals. We continually measure the accuracy of our technology and improve it. Is it biased against skin tones? At Yoti, we take our ethical responsibilities as a company developing new technology very seriously. The data (face image and month and year of birth only) used to train the algorithm is obtained by Yoti in accordance with the UK GDPR during the onboarding process for the Yoti app or using consented data collection exercises. We train our AI with images from a wide demographic of society and then test how well it estimates people’s age for different genders and skin tones. This allows us to see where it needs to improve so we can train the model in those areas. We also invest heavily to minimise bias and make sure it works well for everyone. You can see in our white paper that there’s minimal bias across gender and skin tone for 6-12 year olds, with more bias for older adults. While this is something we’re improving, it’s more important that the model is accurate for 18-25 year olds. Is age estimation secure? We created age estimation to give everybody a secure, private way of proving their age. Security, therefore, isn’t just a priority, it’s fundamental to everything we do. Privacy is a key consideration of age estimation – there’s no login required or the need for an account with Yoti. People simply present their face to a webcam or camera on a device, and as soon as their age has been estimated from that image, the captured image is deleted. We have been independently audited both by the ICO and KPMG on this process of deleting images. The ICO’s report is public and KPMG’s is available on request. We also commission regular external audits of our business and have been certified to meet some of the world’s most stringent security standards, such as ISO27001 and SOC2 Type II. Is my privacy protected? 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 is designed to minimise the data shared, so instead of requesting an ID document, it just needs an image. This image is never seen by a human and is deleted as soon as the age has been estimated. We also designed it so that no individual can be identified by the model. For this reason, the ICO has stated that it can be distinguished from facial recognition technology, as it’s not being used for the purpose of uniquely identifying individuals. Instead, it’s used to categorise people by age. They concluded that our age estimation tool will not result in the processing of special category data. How is age estimation currently being used? Instagram is using the technology to verify the age of users in the US that change their date of birth from under 18 to 18 or over to make sure both teens and adults are in an age-appropriate experience for them. Social network Yubo uses age estimation to check the age of users, ensuring 13-14 years olds are not engaging with adults. UK supermarkets have trialled age estimation at self-checkouts to give shoppers an easy and quick way to prove their age, without needing to show an ID document or wait for assistance. During the trial, shoppers purchasing alcohol could simply look at a camera on the self-checkout to verify their age before completing their purchase. If the system detected a customer looked younger than 25, they could prove their age through the free Yoti app instead. For those who did not wish to use this technology, they still had the option of showing their ID document to a member of supermarket staff. Any other questions? We hope that this answers some of your questions about age estimation. If you still have any questions or would like to learn more about our technology, feel free to get in touch.
Facial Age Estimation white paper
Making it faster and safer to prove your age Our age estimation technology accurately estimates a person’s age by looking at their face. We built it to give everyone a secure and private way of proving how old they are in different everyday scenarios: from age checking on social platforms and online stores, to supermarket self-checkouts, bars and clubs. This privacy-friendly approach to age verification doesn’t require any personal details or documents, and all information is instantly deleted once someone receives their estimated age – nothing is ever viewed by a human. Key takeaways from the report The current accuracy rates (Mean Absolute Errors) are: 2.96 years for 6-70 year olds. 1.52 years for 13-19 year olds. 1.56 years for 6-12 year olds. Users are not individually identifiable. Helps organisations to meet Children’s Codes or Age Appropriate Design Codes. Does not result in the processing of special category data. Gender and skin tone bias minimised. True Positive Rate (TPR) for 13-17 year olds correctly estimated as under 23 is 99.65%. TPR for 6-11 year olds correctly estimated as under 13 is 98.91%. Download the executive summary (updated May 2022)Download the latest white paper (updated May 2022)Download the previous white paper (October 2021)
Safer Internet Day: Helping young people thrive online
Today we’re celebrating Safer Internet Day (SID) – an EU-lead initiative encouraging safer and more responsible use of online technology. On this day, organisations around the world promote the safe and positive use of digital technology for young people. As part of the global landmark event, this year, the UK will explore the theme ‘All fun and games? Exploring respect and relationships online’. Young people are shaping the interactive entertainment spaces they are a part of and SID 2022 celebrates their role in creating a safer internet, whether gaming and creating content or interacting with their friends and peers. This year’s theme is a direct response to the negative behaviour young people are faced with on gaming platforms – bullying, verbally abusive language, groups ‘ganging up’ against other groups, and hate directed at LGBT+ users being just some examples. As well as creating a harmful environment, it also distracts from all the positive aspects that gaming and social platforms have to offer. With age assurance tools like our facial age estimation technology, we know we can help to create more positive experiences for young people using these channels. Here’s how… Providing age-appropriate services to all As of September 2021, the UK’s independent data authority, the Information Commissioner’s Office (ICO), introduced the Age Appropriate Design Code, also known as the Children’s Code. The Code aims to give young people a safe place to play, learn and explore online and is now a legal requirement. It outlines a set of standards that social media, dating, gaming, and other online services likely to be accessed by minors should follow. It also requires businesses to design services to be age-appropriate from the ground up. Before we can safeguard children online, we need to prove they are in fact children in the first place. Our facial age estimation allows us to do exactly this and we believe it will prove a valuable tool going forward in re-shaping the standards for gaming and social platforms! Following the children’s code We’re leading the world in age assurance to enable children to enjoy their online experience. Our work with social media, dating, gaming and other online services help them to make their platforms safer by design and ensure compliance with the Code. We can support platforms to recognise their users are a minor and discourage them from employing nudge techniques and ensure children aren’t asked to provide unnecessary personal data or make their real-time location publicly available. We’ve developed an array of sophisticated age assurance methods that let businesses gain the right level of assurance that someone is the appropriate age. They can choose from a range of age assurance methods to do so, from an estimated age based on a real-time selfie or a verified date-of-birth from an ID document. We only share the result of the age check, and instantly delete all personally identifiable data. Our formidable AI-powered facial age estimation technology allows anyone to be age checked by just looking into a camera on a device. The technology is built to be anonymous and has been proven to be much more accurate than humans, estimating 6-12-year-olds within 1.28 years and 13-19-year-olds within 1.55 years. To discover more about the accuracy rates, read our regularly-updated white paper. Once you’ve discovered that your user is a child, you can then design a different user experience for them in line with the code. Our age assurance tools make it about knowing they’re a child and then providing them with age-appropriate gaming and social experiences, allowing them to thrive in the digital world! Learn more about our age assurance tools.
The UK's Online Safety Bill: moving towards a safer internet
The Online Safety Bill is the UK Government’s chance to make the internet safer for all. It is also a major step forward in the battle against online harms and goes hand in hand with our commitment to protect young people and the vulnerable online. While it’s the responsibility of everyone to make the internet a safer place, regulation is necessary to make businesses act responsibly. The Online Safety Bill’s key recommendations Since a first draft was issued in May, MPs and peers have made four recommendations, as well as many others, to strengthen the Bill: What’s illegal offline should be regulated online Ofcom should issue binding Codes of Practice New criminal offences are needed Keep children safe from accessing pornography We’re heartened to see the proposed revisions include the need for age assurance technology to protect children and the requirement to abide by minimum standards, which we heartily endorse. We are specifically pleased to see that the Bill recommends keeping children safe from accessing pornography: “All statutory requirements on user-to-user services, for both adults and children, should also apply to Internet Society Services likely to be accessed by children, as defined by the Age Appropriate Design Code. This would have many advantages. In particular, it would ensure all pornographic websites would have to prevent children from accessing their content. Many such online services present a threat to children both by allowing them access and by hosting illegal videos of extreme content.” In the UK, it’s estimated that 48% of adolescents have viewed pornography online. We hope that this recommendation in particular will help create a shift in business attitudes which lean towards protecting children within digital spaces. In the Bill, advocate for children’s rights in the digital world, Baroness Kirdon, provided a vital statement on age assurance: “Protecting children is a key objective of the draft Bill and our report. Our children have grown up with the internet and it can bring them many benefits. Too often, though, services are not designed with them in mind. We want all online services likely to be accessed by children to take proportionate steps to protect them. Extreme pornography is particularly prevalent online and far too many children encounter it – often unwittingly”. The Office of the Children’s Commissioner revealed that over half of 11–13-year-olds have seen pornography online. Witnesses explained that pornography can distort children’s understanding of healthy relationships, sex, and consent by, for example, normalising violence during sexual activity. “Privacy-protecting age assurance technologies are part of the solution but are inadequate by themselves. They need to be accompanied by robust requirements to protect children, for example from cross-platform harm, and a mandatory Code of Practice that will set out what is expected. Age assurance, which can include age verification, should be used in a proportionate way and be subject to binding minimum standards to prevent it being used to collect unnecessary data.” What is age assurance? Age assurance describes the methods and measures that help to determine a person’s age or age range. The word assurance refers to the varying levels of certainty that different solutions offer in establishing an age or age range. This includes age verification and age estimation, which both have varying levels of confidence. To learn more about age assurance, read our article about it here. A tried and tested solution for the safeguarding of children We help businesses implement robust age assurance within their services, aligning with the Online Safety Bill and the Age Appropriate Design Code. Platforms can age assess and ensure that children are treated as children – then not serve notifications, avoid tracking, avoid geolocation, avoid inappropriate advertising to minors, provide age appropriate support and content moderation to enable young people to thrive online. To read more about what we’re doing to follow the standards within the Age Appropriate Design Code, you can read more about it here. We work closely with businesses and regulators to help young people thrive online safely. If you would like to learn more about how you can be a part of the evolving online safety landscape, contact us here.
Age assurance: helping young people thrive online
“Age assurance is simply a gateway to the bigger prize of building the digital world young people deserve.” – 5Rights Foundation, But How Do They Know It’s a Child? Recent advancements in regulation and technology have transformed the ways we now check how old someone is online. Until recently, designing age appropriate services for young people from the ground up didn’t seem possible, with businesses accepting that young people would simply have to face being exposed to online harm. But, thanks to intelligent age assurance technology, we now have valuable and powerful tools that can help platforms start designing their services so that young people can be supported to thrive within digital spaces in a way that is privacy-preserving, secure and simple. Age assurance: age verification and age estimation Age assurance describes the methods and measures that help to determine a person’s age or age range. The word assurance refers to the varying levels of certainty that different solutions offer in establishing an age or age range. This includes age verification and age estimation, which both have varying levels of confidence. Age verification refers to the process of age determination by reference to identity attributes, like a passport, that provides evidence of a user being of a certain age. This has traditionally provided the highest level of confidence in a user’s age. However, the level of confidence depends upon a number of factors: for example, a person’s ability to borrow a friend’s ID document, or using fake IDs. On the other hand, age estimation offers the possibility to age check the estimated one billion people globally that don’t own identity documents and can use facial analysis and machine learning techniques. Like age verification, it is also a subset of age assurance, but refers instead to the assessment that someone is likely to fall within a category of ages, or whether they are over or under a certain age. There are many ways that a person’s age can be estimated: Self-declaration Biometrics Profiling and inference models Capacity testing Cross-account authentication Third-party age assurance provider Account holder confirmation Device controls Flagging Well-trained AI models are capable of estimating age more accurately than human judgement. This is because humans aren’t always trained at identifying a person’s age – and also can’t always detect when a document could be fraudulent. Age estimation tech is also scalable – capable of completing thousands of checks a day successfully without getting tired. There is also compelling evidence that certain types of age estimation, such as facial age estimation, can provide a level of confidence in a person’s age at least as high as traditional forms of age verification. So, age estimation could very well provide greater assurance than human checks. What is important is finding the sweet spot that works for your business according to the regulations in place within your industry. The importance of age assurance for young people today Today, young people live out much of their lives digitally. During the pandemic, 1.4 billion young people across 190 countries were not able to attend school in person and much of their learning moved online. “Age assurance should not be mistaken for a silver bullet or a shortcut to making the digital world fit for young people […] Its value lies not simply in the act of assuring age but in the enormous opportunity it brings once young people have been recognised.” – 5Rights Foundation, But How Do They Know It’s a Child? Once businesses are able to recognise that a child is using their service, only then can they devise services fit to support and protect young people. Children have a right to use the internet and digital services as much as adults do, and age assurance is a powerful tool that will empower young people to safely navigate these digital spaces and thrive online. We believe that everyone should be able to prove who they are. But with an estimated one billion people worldwide who don’t own a legally-recognised form of ID, we need solutions that do not rely solely on hard identifiers. Age estimation is hugely beneficial to people across all demographics, particularly to young people. An estimated 24% of over 18s and 33 percent of under 18s in the UK do not have access to photo ID, so other ways must be available for them to prove their age.. Ensuring they are able to prove their age by a variety of means will enable them to navigate the internet safely. How we’ve developed rights-respecting, accessible age assurance Developing our facial age estimation technology requires a data set of facial images where the person’s age is already verified. We’ve trained our algorithm on a wide range of ages, genders and skin tones to minimise bias and have collected data in accordance with the UK GDPR. We’re transparent in how we develop the tool and reach out to organisations to make sure our approaches are in the best interests of our users. We’ve shared our approach in roundtable discussions with industry leaders to gain valuable insight and expertise that helps us develop the tool in the best way possible. This has allowed us to confidently apply facial age estimation to tackle online grooming and launched a #sharetoprotect movement to improve the diversity of the dataset. Meaningful age assurance is no longer just an advantage… Some industries have already recognised the need for age assurance, and so have enforced legislation making it a compulsory practice. For example, UK video sharing platforms must comply with Ofcom’s regulation guidance, and gaming companies should apply the Principles for online and app-based games. But it’s not just businesses delivering age-restricted services that need to worry about age assurance. A number of legislative changes have happened in the past few years, and there are more on the way that could affect the way businesses across the world operate. Any industry where young people might be more inclined to access its digital services would benefit hugely from age assurance measures. It’s also no coincidence that these industries are also the same ones impacted by changing and updating legislation, such as the ICO’s Age Appropriate Design Code (the Code). The Code came into force on 2nd September 2021 and aims to give young people a safe place to play, learn and explore online. With that in mind, if your business operates digitally and is used by young people, it might be time to deliver appropriate age assurance with some of your online services. The bottom line is no one knows when you could be impacted by legislative changes. We think taking a ‘Safety by Design’ approach is the best way to ensure ethical age assurance. “[Safety by design] focuses on the ways technology companies can minimise online threats by anticipating, detecting and eliminating online harms before they occur.” – Julie Inman Grant, Australian eSafety Commissioner This, combined with Privacy by Design principles, will protect a digital service from being potentially harmful to vulnerable people and young people online. Holding each other accountable Our goal is to start a movement of rights-respecting age assurance that upholds community values and protects people’s privacy. That’s why it’s not only important to build trust among our users but to do it in the right way. We do this by having internal and external bodies that hold us accountable—for example, our internal trust & ethics board and our external Guardian’s Council. The Guardian’s Council – made up of four independent advocates – helps Yoti always seek to do the right thing. They provide us with new and different perspectives from key backgrounds like law, tech, data and not-for-profit. We have also set up an internal Ethics & Trust Committee, which includes members from several different areas of our business to consider ethical issues related to our technology and its use. This, alongside a number of public commitments to build ethical technology and our B Corp status, means that we uphold values that benefit the integrity of those who use our services. Read more about what we’re doing to build solutions with trust at the core.
Auckland Transport adopts digital student ID cards
We’re delighted to announce that we have partnered with Auckland Transport to provide digital ID cards for Auckland Secondary school students, providing an innovative alternative to plastic cards and reducing waste. This change is just one of the ways cities like Auckland can make their infrastructure even smarter by using digital-first technologies. The more accessible way to travel The digital ID cards offered through the new scheme will be free of charge to schools. Until now, providing student ID cards was an expensive and time-consuming process. By using a free digital ID app like Yoti, card creation and distribution is much faster. Aine Conlan, Auckland Transport’s Manager of Customer Care says, “Auckland Transport are excited to be able to help provide cost effective easy ID options for the secondary schools children of Tāmaki Makaurau”. Digitising the way we travel helps us to free up resources for schools. And, by producing fewer plastic cards, it also reduces the impact on the planet and local environment. There’s also the issue of lost or stolen cards – we’ve all experienced it. If a student were to lose their phone, their Yoti account and digital ID can be just as easily restored on a new device by them within minutes. What’s more is there’s no added cost, lessening the administrative burden for schools. Aine Conlan, Manager, Customer Care, Auckland Transport: “AT is excited to be partnering with Yoti to provide this new digital ID option for secondary school students. It was important for us to be able to work with an organisation that could offer our customers the security and flexibility that they need”. As part of the scheme, physical cards will still be available for any student who may simply prefer using them or perhaps doesn’t have access to a smartphone. Once the student is set up with an account, they will have a free, reusable ID that can be used in a growing landscape of businesses adopting digital ID. Karl Stemp, Commercial Director, Yoti NZ: “We are excited to be offering Digital ID to all Auckland secondary schools and students. Auckland Transport adopting this innovation is showing great initiative to make the concessions process more customer focused and giving people a plastic free option for ID cards makes sense as part of these improvements”. Privacy matters We’ve adopted a data-minimised approach to identity. That’s because we believe people shouldn’t have to share everything about themselves when completing day-to-day tasks like getting the bus. And with our app, you can share less data. We’re not in the business of selling or passing on data to any organisations – only the owner of the ID can control what data is shared and when. All data is securely encrypted, so we don’t have access to anyone’s personal information. If you want to learn more about our digital ID solution, you can read about it here. Alternatively, if you want to find out how digital ID can be used to transform your business or institution in New Zealand, get in touch with our Commercial Director Karl Stemp.