“Some organisations use Artificial Intelligence (AI) to blur personal information within their archives. Others use AI to retrieve ID information from scanned passports and ID cards, in order to enrich their databases.” This month in AI on the road, Krijn talks about the dual use of AI in processing passports and ID cards.
In the column “AI on the road”, Hyarchis Artificial Intelligence expert Krijn talks about the AI-related questions that he encounters during meetings with clients. In other words, very basic questions and answers about this seemingly complex topic. This month Krijn focuses on the use of AI for managing sensitive passport- and ID-card-related data within large archives.
Krijn, let’s start with using AI to remove ID information. What questions do you encounter?
The questions that arise around removing ID information, or better said: making that information inaccessible, are often based on compliance with GDPR legislation. Various organisations are burdened with large archives containing millions of copies of passports and ID cards. Under the current regulation, it is important that access to sensitive data, such as personal identification numbers (PINs), is restricted. AI offers the possibility to do this while leaving the integrity of the archive intact.
The Dutch website InFinance recently published an article on the subject, in which they outlined how to correctly process PINs in line with the regulations of the Dutch Data Protection Authority (Dutch: Autoriteit Persoonsgegevens). AI gives organisations the ability to track down personal identification numbers in documents which are not text-searchable. As a result, organisations can find and subsequently hide sensitive data in an automated manner even in outdated or complex archives.
Could you give an example of information that organisations need to blur?
Before GDPR came in effect, the PIN of someone who was applying for example a mortgage was recorded and saved in this file. Under current legislation, PINs are not to be disclosed in a this application, unless this is essential to the purpose for which the data is processed. Only when the contract is actually signed, the personal identification number can be recorded and included in the case file. Because of this, many organisations are now wondering “How do I make this number inaccessible?”
What role does AI play in all of this?
It is impossible to manually remove specific information from scanned passports and ID cards. Passports might be scanned horizontally, vertically, crooked, or two scans could be even combined on one page. Add the wide variety of layouts – every country has their own passport design – and you start to understand the magnitude of this challenge.
Step one is automation. However, what do you do if not only the search process has many variables, but the blurring process too? That is where AI comes into play. Using OCR-based technology, our software solutions recognise personal identification numbers within scanned documents with a near 100 percent accuracy. Once found, AI makes sure that these numbers are blurred, making them unrecognisable.
What is the difficulty that you address by using AI?
It is often unknown where the files containing passports and ID cards are located. Most of the time the files are not named accordingly and located somewhere in a 50-page personal file. AI traces these regardless of the file type or file name and ensures that the PINs are blurred and no longer accessible.
If I understood correctly, AI is not only used to remove information from digital files, but also to add information to the database? Tell us more…
Indeed, clients also use AI to extend their database. In the past, passports and ID cards were not always registered properly, because there was no need to do so. Nowadays, organisations do need to implement appropriate data minimization measures and store personal information transparently for the customers in order to serve them better and comply with data privacy laws and regulations.
Now I’m curious…is that possible?
Yes, it is possible, but to do that, you need to use AI. If an organisation would like to recognise the data on an ID card or passport that is in their archive, our AI solutions can help with that. Whether you want to remove existing information from your archive or do the exact opposite, our AI-based software works both ways!