About Governance Policy Registry

Artificial Intelligence

AI testing for age estimation

Used to mimic human cognitive functions, artificial intelligence (AI) and machine learning technologies are becoming increasingly prominent and advanced. In the age restricted services and goods field, they are now being used to predict age through just an image taken by a camera - often without referencing other ID or data sources. 

But, not even computers can be 100% accurate, 100% of the time. So, we test and validate the algorithms deployed by AI and machine learning technologies to make automated judgements about the age of customers to ensure they’re accurate and unbiased.

We apply ACCS 1:2020 - Technical Requirements for Age Estimation Technologies.

Any company looking to deploy or introduce AI age estimation technology in the UK needs this certification to verify that their solution works accurately within the acceptable age tolerances (such as challenge 18 or 25) and is free from gender or racial biases.

Why would I use AI to test age?

Age verification without referencing IDs or data sources is useful when there are sensitivities relating to the service or product on offer, or if there is need for an on-the-spot assessment of the subject’s age. For example, users may be uncomfortable sharing their personal information on adult entertainment websites, due to data privacy or breach issues, so the technology simply verifies their age by scanning their face, without collecting any other information from them. 

These technologies also work well in retail and convenience shops where there is need for a quick check out process. The technology can be used via an app or integrated into self-service vending equipment for bars selling alcohol or into slot machines for gambling shops.  The technology can also be deployed in CCTV systems used by law enforcement for surveillance.

Why is testing it so important?

  • Ensures accuracy

Our testing protocols use statistical analysis in combination with other methods to check the accuracy of algorithms used for any age estimation technology. AI solution providers usually claim accuracy between 95-99%, but this is reliant on the sophistication of the technology used and the data training sets used to develop the algorithms for the AI. Our job is principally to test this claim of accuracy made by the client.

  • Instils customer confidence

Third party testing of AI age estimation technology is also important because it provides assurance that the system has been independently verified by a reputable organisation and the claims of performance are valid and reliable, helping to instil customer confidence and reduce claims of inaccuracy.

  • Fights bias

Our assessments can also identify inherent bias that may lie unknown within these technologies. AI algorithms could have inherent bias to the recognition of individuals with different skin tones or genders, which subsequently leads to discrimination and disenfranchisement of certain groups who may use the technology.

  • Ensures compliance

Ours is the only certification recognized and approved by the UK Home Office and the Equalities and Human Rights Commission for validation of this technology.


Our state-of-the-art app enables our test purchasers to capture live, geo- coded, time stamped results providing instant feedback on age check performance.

Request a Demo

We have hundreds of vetted, trained test purchasers available to undertake visits to client premises.

Find Out More

We use cookies to offer a better user experience. The cookies enable you to enjoy certain features and are asked for performance, social media and advertising purposes. By continuing to use this site, you accept our Cookie and Privacy Policy.