How can Identity Matching improve the experience of citizens on online public services?
©Pixabay
There is a growing shift from paper-based documents and procedures towards electronic records and online services in Public Administrations of the European Union. This transformation means that large number of databases are getting connected to each other in order to exchange relevant information between the administrations.
However, those databases are often tailored to the service provided by each administration. Different information is collected about citizens and businesses and may identify people and organisation differently: the social security department may use a national social security number or an association of the name and date of birth of a person, while the taxation office may use a dedicated national tax number, a VAT number, etc.. In addition, differences in the spelling of the names, cities and different approaches with regard to the record of second names, married and maiden names add to this complexity, leading to a situation in which the same person is not always represented under the same identity data.
This means that when databases are connected, it may be difficult for an administration to match the identity of a person in one database to their record in another database. This situation, which is already complex at the national level, is now amplified by the interconnection of the EU-28 Member States' databases, which leads to a real challenge: how to perform identity matching in order to ensure citizens and businesses a smooth experience when accessing online public services accross-borders?
In recent years, EU Member States have been working hard to enable citizens and businesses to access online public services using their national electronic identification solution. These efforts fall directly under the scope of the eIDAS Regulation (EU) 910/2014 on electronic identification, authentication and trust services, which aims at making national eID schemes interoperable across Europe in order to facilitate access to online public services.
As a result of the eIDAS regulation, citizens and businesses are progressively able to identify and authenticate themselves with credentials they use in their own country on foreign online public services. However, if the service they want to access cannot match their identity to a record, the experience of the citizens and businesses may be disrupted. For instance, the access to the service may be declined, there might limitation of the service, or there could be the need to shift back to paper-based solutions as a fall-back option, as the citizen would not be recognized as a person for which a previous record already exists.
Conscious that this problem is posing a real barrier to the smooth access to online public services across borders, Member States are working together to exchange best practices on how to perform identity matching following an eIDAS cross-border authentication.
In Spring 2019, EU Member States have already organised two sharing sessions to discuss the topic of identity matching.
Member States discussed the following challenges among others:
How to perform linguistic based matching ?
Matching names from different contexts is not easy due to cultural differences concerning spelling and phonetics. For instance, some names may be transliterated from different alphabets. Whilst the process is easier for certain alphabets, e.g. one Greek character is equivalent to one Latin character, the situation becomes more complex with alphabets such as Cyrillic or Arabic. Regarding the former, the name Борис Ельцин would be known as Boris Yeltsin in English, yet it would be written Eltsine in French or Jelzin in German; whilst in Arabic, the name محمد can be written in Latin characters in many different ways, namely Muhamad, Mahomet or Mohamad.
As a result, central governments and private relying parties are increasingly developing methods to ensure a reliable quality in matching names. With references to the examples mentioned above, if the spellings come to be matched against each other, statistics would align only the most similar ones. Dictionaries could store the most frequent variants and reduce the probability of missing a hit, but would probably be far from being exhaustive. In order to guarantee a match in all cases, a “back to the roots” approach is used, first reconstituting the original script, i.e. in this case in Arabic, then using the internal transliteration mechanism to get the same Latin references for matching.
- How to address false positive?
False positive consists in creating links between identities that do not correspond to the same person.
Initially, simple algorithms were developed to handle name matching relying on statistics to identify similarities. These simple approaches failed to produce correct results and often resulted in many false positives in matching two identities. This is especially the case with the phonetics of certain names, which produce similar characters but can be spelt in a variety of different ways. For example, one of the most common surnames in German-speaking countries is Maier, which could also be spelt Mayr, Maier, Meir and Meyer. Due to the magnitude of different matching results there is thus the probability of causing wrong connections, by matching a user to somebody else’s identity.
Therefore, the current trend of innovation is implementing artificial intelligence techniques and in particular machine-learning algorithms to reduce the number of false positives and allow for auto-resolution. This approach has been adopted by some Member States, where probability calculations are made to reduce the chances of false positives in the national database.
Member States are facing similar challenges linked to the digitalisation of their public services and national records. The European Commission will therefore continue to support them by facilitating the exchange of information and best practices between EU Member States in the upcoming months.
The Connecting Europe Facility (CEF) eID Building Block primarily supports the Member States in the roll-out of the eIDAS Network (the technical infrastructure which connects national eID schemes). CEF eID is a set of services (including software, documentation, training and support) provided by the European Commission and endorsed by the Member States, which helps public administrations and private Service Providers to extend the use of their online services to citizens from other European countries.