Rabat – A recent report by Lighthouse Reports has shed light on the so-called “information-supported decision-making” used by the Dutch government to process short-term visa applications to the country and the Schengen area, revealing that it encourages racial bias and discrimination.
The findings raise many questions and concerns: how does the algorithm work, and for how long has it been operating? What data does it use? What impact does it have on visa applicants and their rights?
Whistleblowers have disclosed that the Dutch ministry’s algorithm potentially encourages discrimination and profiling of short-term visa applications.
For instance, an applicant from Morocco or Suriname would be more scrutinized based on the algorithm since the software filters out applicants based on nationality, gender, age, and place of application. The revelation suggests that the algorithm may perpetuate biases and hinder fair decision-making.
Furthermore, the investigation showed that the new digitized process has resulted in longer queuing times, accessibility issues, and delays for prospective applicants, especially travelers from certain countries, such as Morocco.
According to the Lighthouse Report, the Dutch Ministry of Foreign Affairs has been using this algorithm without the knowledge of the public to process short-term visa applications to the Netherlands and the Schengen area since 2015.
The visa application is mandatory for all non-EU citizens intending to visit the country for tourism, family visits, or business purposes.
VFS Global, a Dubai-based multinational company that handles visa applications for the Netherlands, gathers documents from applicants, computes the applications, and sends them to the Dutch Foreign Affairs Ministry for eligibility screening.
Using the algorithm, officials can “save time” since the software filters out low-risk and high-risk applications based on various factors, including the applicant’s nationality, which receives more scrutiny.
Despite warnings from the Dutch data protection authority and promises from Dutch Foreign Minister Wopke Hoekstra to investigate this issue, leaked documents obtained by Lighthouse indicate that the ministry kept using the discriminatory algorithm, disregarding the raised concerns.
In an interview with Morocco World News (MWN), Amin Byad, a data protection expert based in Casablanca, highlighted the issues associated with profiling software like the one employed by the Dutch Ministry of Foreign Affairs.
Byad explained that using such profiling algorithms may lead to discrimination, bias, and errors in the visa process, potentially undermining an applicant’s chance of getting the visa.
He emphasized that because these algorithms collect data about people, they may “violate the principles of data protection law, such as fairness, transparency, accuracy, and purpose limitation.”
“Using this algorithm/profiling software is a breach of data confidentiality/privacy because it involves processing the personal data of visa applicants without their consent or knowledge,” he added.
Byad argued that using such software should be closely regulated as it leads to ethnic discrimination “because it may rely on stereotypes or assumptions based on the applicants’ nationality, gender, age, or place of application.”
For technology lawyer Wouter Seinen, the concern lies not with the algorithm itself but with the selection of nationality as a parameter to determine an application’s level of scrutiny.
“The trigger is clear; it certainly leads to discrimination. Yet, what is unclear to me in the report is whether it is justified or not justified because countries have the right to make policy choices,” the Netherlands-based lawyer, who works for Pinsent Mason, told MWN.
Seinen also suggests that if the policy to subject applicants from certain countries to increased scrutiny was already established prior to the introduction of the software, it might be legally justified, albeit discriminatory.
“In that case, the software is doing exactly the same as the civil servants did before the software took over a bit of that task,” he explained.
Indeed, the Lighthouse report shed light on the pervasiveness of systemic discrimination within the Netherlands’ institutions.
The report highlighted the deeply ingrained practices within the institutions that perpetuated discrimination and unfair treatment, with the Ministry of Foreign Affairs being just one example of Dutch institutions that have faced scrutiny for their role in perpetuating institutional racism.
“If a government body is doing the right thing, they should constantly challenge their software and check whether the outcomes are reasonable,” said Seinen.
Real-life implications
Saadia Ourhirs, a Moroccan Dutch mother, experienced this discrimination firsthand in 2020 when her aunt’s visa application to support her during childbirth was repeatedly denied.
Saadia started the visa process for her aunt in the summer of 2020, hoping she could come to the Netherlands in the spring of 2021 in time for her niece’s birth.
Yet, the lengthy visa process and difficulty of securing an appointment with VFS Global made that impossible.
Apart from the discriminatory implications, the investigation uncovered inadequate security measures in the Ministry of Foreign Affairs system used for the Schengen Visa process. This vulnerability places applicant data at risk of security breaches, further compounding data protection and privacy concerns.
In February 2020, the Dutch privacy regulator fined the Dutch Ministry of Foreign Affairs €565,000 for “inadequately securing visa applications” and violating the European Data Protection Regulation (GDPR).
The key finding from the investigation highlighted that the Ministry of Foreign Affairs system used for the Schengen Visa was not adequately secured, which put applicants’ data at risk of security breaches.
It is also not the Netherlands’ first time facing an algorithm-related scandal.
Back in 2019, the Dutch tax authorities stripped several poor families from child benefits using a self-learning algorithm to spot childcare benefit fraud. The families from minority and low-income backgrounds struggled in the aftermath of the incident as they were obliged to pay back exorbitant amounts of debt to the agency.
Following the scandal, which wrongly accused many families and stripped their children of needed support, international organizations such as Amnesty International rang the alarm to ban racist algorithms.
The Dutch privacy regulator subsequently fined the tax authority €3.7 million on charges of violating several GDPR rules, including unlawfully processing people’s personal data.
Far from being an isolated case, the Netherlands is not the only country using profiling software and questionable algorithms to streamline public services.
The case of Schufa in Germany illustrates another instance where profiling procedures violated data protection provisions outlined in the GDPR.
In Germany, Schufa is essential in granting access to housing and credit card application approvals, as landlords and banks rely on the company’s credit assessments.
An investigation by OpenSCHUFA, Spiegel Online, and BR raised concerns regarding the fairness of the scoring procedures used by the company, particularly its impact on individuals from immigrant and low-income backgrounds.
Similar issues have arisen with Austria’s employment agency using profiling software to determine job opportunities for unemployed people, which came under scrutiny for gender-based discrimination.
Balancing Technology and ethics
As society increasingly relies on technology, it becomes crucial to address ethical and moral implications to ensure fair treatment to all.
Byad emphasizes that data protection laws can safeguard citizens from discriminatory risks and can set guidelines to ensure equal and fair treatment for all.
“They can also empower citizens to exercise their rights and control over their data. Furthermore, they can create mechanisms for oversight and enforcement by independent authorities or courts,” he added.
The recent findings regarding the Dutch government’s visa processing algorithm have once again put a spotlight on the potential for bias and discrimination in automated decision-making systems.
“We should always be reviewing the effectiveness, the fairness of the processes and not just getting lazy and relying on the computer without interrogating it,” highlighted Seinen.
Amid the rise of AI-powered technology, it is imperative for governments and institutions to ensure the protection of individual rights, prevent algorithmic biases, and uphold fairness and transparency throughout their processes.
Now more than ever, this discussion is important to ensure technology doesn’t become a tool to further discriminate against marginalized groups.
