How we talk when we talk about age assurance

The lexicon of age assurance is a complex stew of technical, regulatory, political and promotional language. Competing agendas, regional differences and legislation all play a role in shaping how we talk when we talk about age assurance. For the average internet user, who just wants to access a site, the language can be a major challenge to navigate. This, despite age assurance being a consumer-facing product.

Managing the language of age assurance is a shared responsibility that requires as much attention as technical implementation and regulatory compliance. The core issue is trust. Constantly changing terminology can create uncertainty, which can translate into risk.

Imagine a pond, on which kids go out to play “shinny” – a strange word that, for whatever reason, Canadians have chosen to mean “a pickup game of hockey played on a frozen lake or pond.” The key word is “frozen.” Unstable ice is a risk, and kids know not to go out on it unless they want to fall through.

Now imagine that pond is made of the lexicon of age assurance – all the terminology used to describe, sell and regulate age check technology. At present, it appears unstable. Innovation and solidity are in tension, as firms look to expand their offerings to service a rapidly evolving market. Some words that don’t belong inevitably get thrown into the mix.

Assure, verify, estimate, infer, check: the key verbs 

To assure is to indicate that something is true, often with a positive slant intended to instill confidence. A restaurant server might say to a customer, “ I can assure you that the beef is very good tonight.” Verifying that claim would require taking a bite. On a more concrete level, in the context of accounting, assurance means providing an independent and professional opinion in order to reduce risk associated with incorrect data.

Estimation is a guess, based on select information. When you get an estimate from a contractor for work, it is an educated guess based on the scope of the job. It may not be exact, and does not have to be. Biometric age estimation guesses a person’s age range based on their facial geometry or other biological characteristics.

Inference is similar to estimation, in that it involves consulting available information to arrive at a conclusion. In age assurance, it typically considers behavioral factors; for instance, by looking at which public accounts are tied to an email address, one might infer from an association with a mortgage broker or a credit card company that a person is over 18. Machine learning-based inference, increasingly used by large platforms, attempts to infer a person’s age based on account information, such as what they watch, who they’re connected to and other available data. The UK Information Commissioner’s Office (ICO) has already differentiated this practice from inference, labeling it “profiling.”

Stable language establishes trust

Friction is the term we use for mechanical barriers to access: a cumbersome password prompt, a redirect. Language can create the same effect in users – a desire to abandon the task at hand – by creating dissonance. If friction is mechanical, dissonance is its psychological counterpart.  It occurs when words don’t quite add up.

Some big companies have taken to labeling their age assurance systems as “age prediction.” However, the word prediction implies a future date: it is foretelling. One might predict a stock market crash or the results of a basketball game, but this must be done before the event occurs. As such, to “predict” someone’s age means to guess how old they will be in the future. Semantically, it does not refer to a person’s age in the present moment – and as such should not apply to age assurance systems that need to know how old their users are at point of access.

Stable language is a shared responsibility. Trust develops in consistency, so it is in the industry’s best interest to try and collectively agree on consistent language in order to build trust. The desire to stand out in a rapidly evolving sector can trigger linguistic missteps that threaten stability. While every company wants to be something new, at some point you must be what you are: one can dress up water by saying it comes from a spring, that it’s mineralized or enhanced, or branding it as “Liquid Death” – but it is still water. A company attempting to sell water by calling it “clarity juice” or “lake sweat,” for instance, is unlikely to win many customers, and could well catch hell from regulators.

Stable, consistent language is tied to transparency. As age checks have hit the global headlines, awareness of age assurance companies and practices has spread. While public opinion tends to support online safety, distrust of individual companies remains a problem. Witness the case of Persona and Discord, wherein a keen Redditor combed the company’s privacy policy and offered criticism – which eventually led to Discord severing ties with the provider. Privacy policies are fine print, but they are not hidden: resourceful digital natives will find them, and it only takes one person to highlight an inconsistency that can collapse wider trust.

Supporting standards with timely updates 

The terms “age assurance” and “age verification” are still routinely conflated, for various reasons. But at some point, that must stop, since they are not the same thing.

The new ISO/IEC international standard on age assurance, 27566-1, attempts to lock down definitions so they may better serve as a global reference point. In different languages, the words themselves may differ, but the concepts they express are established and agreed upon. Ultimately, the standard describes core characteristics for enabling age-related eligibility decisions. It is not a technical specification, but a guide to conceptual stability.

That said, standards move much more slowly than innovation, particularly in the age assurance sector at this point in time. A good illustration of this is the international standards for presentation attack detection – a term now typically included under the more recent umbrella term, “liveness detection,” which has not yet made its way into the standards.

ISO/IEC 27566-1 will not keep pace with changes in the industry. Updates will occur on a predetermined schedule, but bureaucratic processes will never match the speed at which the tech industry moves. As such, there is a need for independent third-party sources of information that operate in tandem with rapid changes in the sector. In keeping tabs on how language is shifting, one can ensure to avoid gaps in the ice, and better contribute to the shared job of maintaining solid ground on which the age assurance industry can thrive.

*This article is based on a presentation given at the 2026 Global Age Assurance Standards Summit in Manchester. 

Article Topics

age assurance standard  |  Age Assurance Standards Summit (2026)  |  age verification  |  biometrics  |  digital ID  |  facial age estimation (FAE)