Deceptive looks: can machines pinpoint person's age?

Detecting a person’s age online is becoming faster, more accurate, and surprisingly smart. But is a selfie really enough to determine someone’s age? What role does biometrics play in combating various types of fraud? Aleksander Tsuiman, Head of Product Legal and Privacy at Veriff, provided insights in an interview into a world where a six-second decision determines whether you can access a social network, buy alcohol, play video games, or confirm if the person in front of the camera is indeed the one holding the passport or ID card.

The debate on age restrictions for minors

Australia has taken a step towards restricting social media access for minors due to concerns about its impact on their developing worldview. Similar debates are taking place in several U.S. states. While a store clerk or security guard might be able to estimate a person’s age at a glance, it’s not foolproof. Sometimes, a youthful-looking adult is mistaken for a minor, while a teenager with newly grown facial hair can slip past age filters.

So, the obvious question arises: could machines do this better than humans?

Where Veriff comes into play

Veriff skyrocketed in popularity by offering identity verification using video camera images. The user shows their ID in front of the camera, and Veriff’s platform cross-checks the visual data with information from the document. The system also includes a "liveness" test to confirm that a real, live person is present — all part of a process that takes just six seconds.

But what happens if an ID isn’t available, or if requesting it would reveal too much personal information? Aleksander Tsuiman explains the company’s approach to this challenge.

Using biometric data to estimate age

Tsuiman confirms that online age detection is entirely possible. AI can analyze our features well enough to estimate our age with surprising accuracy. But what are the most effective technologies for this? Does it rely on documents, facial recognition, or something else?

The general trend is toward using biometric data for both identity verification and age estimation. However, a combination of different checks and data points is more accurate and secure than relying solely on facial recognition. For example, if biometric data is supplemented with ID document checks, the likelihood of accurately determining a person’s age increases significantly.

Veriff’s technological developments

Veriff continues to expand its portfolio of biometric verification tools. Tsuiman notes that some exciting developments are in the works, and more details may be revealed next year. Currently, Veriff employs Biometric Authentication for repeat users who have already undergone a comprehensive identity verification process.

"While most people are honest, around 5-6% of all sessions involve fraudsters attempting to pose as someone else," says Tsuiman.

The company's verification process analyzes more than 1,000 different data points, including biometric and behavioral data, document information, and device and network details. If necessary, additional verification checks can be added upon a customer's request.

The process is largely automated, with decisions typically made in about six seconds. If the automated system cannot reach a definitive conclusion, human reviewers step in, which may take more time. The demand for real-time responses has driven improvements in the service, reducing decision times from several seconds to just one second for repeat users undergoing biometric authentication.

Veriff is also developing an Age Estimation product. It’s a solution designed to enable rapid age verification with just a selfie — no ID required.

This tool is particularly useful for companies that want to ensure user safety (especially for minors) but prefer not to request ID documents due to privacy or ethical concerns. The system can be used in both virtual and physical settings, such as video game platforms, social media sites, dating apps, self-checkout purchases of age-restricted products, or entry to bars and nightclubs.

According to Tsuiman, Veriff also uses cross-session data analysis for fraud detection. This means that data from multiple verification sessions is analyzed to identify so-called "serial fraudsters."

While most users are honest, about 5-6% of all sessions are flagged for fraudulent attempts to impersonate someone else (source: Veriff Identity Fraud Report 2025)

How does age verification ensure privacy and data security?

User privacy and security are paramount. People need to know why their data is being processed, who has access to it, and how it will be used.

Any form of personal data processing, including data collection, must have a lawful basis. For example, in most cases, the platform (like a social media network) is responsible for ensuring legal compliance since they control how the data is used.

Data security depends on the measures in place to protect confidentiality, integrity, and availability. Encryption during data transfer and storage, controlled internal access to data, and regular system testing are critical measures. Security protocols, policies, and employee awareness are also essential for protecting sensitive information.

Tsuiman notes that companies with industry-recognized security certifications (like SOC2, ISO 27001, etc.) and clearly defined information security principles are generally more trustworthy.

Adapting to New Regulations and International Markets

If a country, such as Australia, decides to enforce stricter age verification on well-known platforms, how easily can Veriff’s solutions be integrated? Tsuiman believes Veriff’s technology would be beneficial for platforms in Australia and in several U.S. states with similar initiatives.

Veriff has discussed potential collaboration with several well-known platforms, but, in most cases, platforms only approach identity verification providers when required by legislation.

One positive example of proactive implementation is Veriff’s collaboration with Starship Technologies. Customers using Starship’s delivery robots must undergo age verification not only when placing an order for age-restricted products but also when retrieving the order from the robot.

How can we prevent the misuse of age estimation systems?

Preventing misuse requires verifying the source of the information. Veriff uses a combination of over 1,000 different data points for identity verification, including biometric, behavioral, and device data, as well as network connection details.

Combining multiple verification methods is always more effective than relying on a single source of information.

Legal and ethical considerations of age verification

Under European Union law, the right to personal data protection is a fundamental right. While GDPR is a key part of this framework, it doesn't prevent states from requiring data collection for specific legal purposes. For instance, health services are legally required to collect certain personal data to provide care, and financial transactions are monitored for audit and regulatory compliance.

When it comes to age verification, companies must assess whether the data collection is necessary and proportionate to achieve a specific purpose. The GDPR principles require data minimization, accuracy, and storage limitations.

If a country mandates age verification, the law should specify the data collection requirements and the type of data that can and cannot be collected. Without clear guidelines, companies must decide on the most effective way to achieve age verification while respecting users' privacy rights.

Avoiding discrimination and overburdening users

Technology must be properly designed and implemented to avoid discrimination against specific user groups. This is especially relevant for platforms with users from diverse regions, whether it's Australia, Estonia, or India.

To reduce the burden on users, companies can design simple, user-friendly experiences. For example, facial recognition-based systems are convenient but may face legal challenges if banned for age verification.

Tsuiman emphasizes that privacy-friendly technologies and solutions exist. Lawmakers should involve industry experts in discussions about regulatory requirements, as this ensures balanced and technically feasible solutions.

How to verify age in countries without standardized IDs?

Many countries have ID systems, but the quality varies. Facial recognition-based solutions offer a practical approach, especially in countries without standardized documents. Where possible, additional cross-checks can be conducted using databases or registries, though this approach raises data protection challenges.

The future of online age verification

AI is already proving its potential to assess age, and future developments will only improve accuracy. Countries like Australia, and some U.S. states, are exploring regulatory frameworks to protect children online. Veriff’s solutions, like Age Estimation, offer a privacy-conscious approach to age verification without needing official ID documents.

With solutions like Veriff’s in development, the ability to accurately verify age using only a selfie may soon become a standard feature of online platforms. It’s a step forward in online safety, privacy, and fraud prevention.