Confidence in Technology

Biometric recognition matches individuals with previously provided biological information, such as eye scans, fingerprints, palm prints, facial features or voice recognition. In our industry, biometrics are used in relation with smart cards or other secure portable devices to grant or to deny access or rights. The trend of the last few years has been to automate controls based on biometrics. This automation (such as in enterprise access control, or in automated border control) is dictated by the need to achieve a high throughput at the lowest possible human cost.

However, biometric technologies are not an exact science, they are also “inherently faillible”. Searching biometric data of an individual in a database does not give a 0 or 1 result, but rather a probability that the person biometrics fits with the record in the database. Setting up a biometric system often mean a careful adjustment of correlation parameters to limit the number of false positives and the number of false negatives according to the population and what is at stake behind the control system.

Then when false positives and false negatives appear, the last resort is human control. And humans are fallible when it comes to recognizing an individual, not only because of biological aspects such as the evolution of the human body over time, but also due to psychological parameters that may affect the decision of the controller. Over time, controllers are increasingly confident in an automated system, and it takes more and more guts for a controller to make a decision contrary to the decision made by the automatic biometric control system. Also, being humans, controllers may be subject to bribery or at least be under influence.

So at the end of the day, we implement a technology that is not perfect, and have it seconded by human decisions that are inherently fallible!

Thierry Spanjaard

Chief Editor

Smart Insights