E-government in the welfare state: Human rights implications of the digitalisation of social policy in Poland
For many years the Polish authorities have been developing large-scale IT projects, and digitising various areas of public administration. These activities were part of realising the e‑government concept in Poland. In broader terms, e‑government means the public sector’s use of information and communications technologies (ICTs) with the aim of improving information and service delivery, encouraging citizen participation in the decision-making process, and making government more accountable, transparent and effective. 1
One of the many sectors where big digital projects were carried out was welfare policy. In Poland – as in many countries – this area is associated with numerous problems (described below). There was an assumption that new technologies would have a favourable effect on the availability of certain services and help to improve the functioning of the administration dealing with social policy. However, there are many concerns as to whether the practical implementation of e‑government could achieve those aims.
Below we will present two examples of the e‑government project conducted by the Polish Ministry of Family, Labour and Social Policy (Ministerstwo Rodziny Pracy i Polityki Społecznej, MRPiPS). These cases will help to describe the relations between social rights (the right to social assistance and right to work) with the sphere of e‑government (which includes both online and offline tools). In addition, we observe that through processing personal information, these systems show the interdependence and interrelation between socioeconomic rights and the rights to privacy and data protection.
Policy background: e‑government and social policy
The development of e‑government in Poland meets many barriers. It is worth noting that Poles – compared to other Europeans – use e‑services less. This is especially noticeable in the area of social assistance (10 times less usage). 2 This is due to many factors. Polish citizens still prefer to take care of administrative matters by visiting offices. Statistics indicate that the most important reasons for not using e‑services 3 in Poland were (among others) the concerns about protection and security of personal data (6%), lack of knowledge and skills (4%), the problem with the e‑signature (2%), and limitations associated with access to websites (1%). 4
Other issues affecting the use of e-government solutions are bad management, inadequate use of European Union funds, and corruption.5 These last two are the core of the political dispute related to the concept of e‑government in Poland.
The digitalisation of social policy was designed to be one of the important solutions to many problems that appear in this area. It is beyond this report to describe them precisely. However, we will try to signal those which are the most serious and set the discussion in this area.
One of them is the growing group of people living in extreme poverty. 6 The issue of poverty is unfortunately not accompanied by effective and commonly accessible forms of assistance and social security. What is also alarming is the small number of people who are granted assistance – at least 40% of extremely poor people do not receive the financial assistance to which they are entitled (due to lack of competencies, knowledge and administrative problems). Social welfare institutions are also underfunded and function in bad organisational conditions. There are also numerous problems connected with the labour market and employment policy: the growing number of people unemployed for a long time and the low share of women in the labour market. For many years Polish authorities also undertook insufficient activities as regards family policy.
Human rights standards
In the context of our two cases we focused on human rights such as the right to social assistance, the right to work, freedom from discrimination and the right to data protection. All of these rights are included in the Polish constitution and various international agreements to which Poland is a party. It is also worth noting that each of these rights is accompanied by a series of specific guarantees stated in statutory laws. 7
Poland is a state-party of the International Covenant on Economic, Social and Cultural Rights (ICESCR). Article 9 of the ICESCR establishes the right to social security and social insurance. Meanwhile, Article 6 Section 1 obliges governments to fully execute the right of a human to work, i.e. through ensuring programmes of technical and vocational consulting and training. The realisation of these two rights should be done progressively while using the maximum of available resources. For measuring the level of fulfilment of obligations relating to these rights, criteria like availability, adequacy or affordability should be taken into account. 8
When performing in practice the right to work and right to social assistance, states also abide by the principle of non-discrimination (which has a status of core obligation). 9 The ICESCR allows the application of differential treatment as long as such measures lead to an improvement of an unwanted situation, are of a temporary nature, and are consistent with the principle of necessity and proportionality. It is also worth mentioning that according to the ICESCR standards, the rules determining who can obtain certain forms of assistance should be justified, proportionate and transparent. Limitations of access to such resources should be based on precise legal provisions and grounds that are reasonable and subject to due process. 10
The Emp@thy project
The Emp@thy (in Polish Emp@tia) project was one of the major IT projects carried out by the MRPiPS. 11 The project started in 2009, and was completed in 2014. The implementation of new IT tools required amendments to the legislation on social assistance – especially in the context of the collection and processing of personal data.
The main objective of this project was to improve the social assistance administration by introducing new technological solutions. It was designed primarily to speed up decisions by officials, unify administrative standards, and ultimately increase the availability of social assistance services. The project also provided more reliable and accurate information about citizens benefiting from the support. Another goal was associated with the rationalisation of the management of public funds – computerisation is expected to bring significant savings in handling the administrative burden. 12
Digitalisation of social assistance in Poland consisted of a series of activities involving products of various types (both online and offline). Among them were:
Central database of beneficiaries – the database collecting information on citizens benefiting from social support and the members of their families. Information from the central register is used to verify the data of people applying for family benefits, alimony funds or social assistance.
Integration platform – a tool for data exchange between public databases, e.g. social insurance, labour offices, tax offices. Thanks to this tool, when applying for social assistance, citizens do not have to obtain various certificates from other administration offices. The data are verified through an online system. This system is also used to detect welfare fraud through analysing data and tracking any inconsistencies.
Analysis and reporting platform – knowledge base of the entire system. It allows the collection of statistical information from the entire area of social security. The platform is designed to help make strategic decisions regarding the distribution of funds allocated for assistance to the communities where it is needed. Another possibility is to detect abuses by people attempting to obtain services in an unauthorised manner.
Information and service portal or simply Emp@thy portal – consists of two sections, one for citizens and the other for officials. The first one is primarily to provide information on the welfare system (the amount of benefits, the rules for granting them, the addresses to local offices). Another feature is the ability to submit an application for social assistance benefits via the web (e‑form). The office will also notify citizens of their decision online. To submit an online application, citizens should log in using an e-signature or via the ePUAP platform. 13 The second section of the portal is aimed at government officials. For example, it allows them to generate statistical data (such as the types of social services most popular in a specific area and demographic data about citizens receiving assistance) and contains information about the interpretation of the rules of the welfare system.
When the project was finally put into use it caused quite a controversy. A series of media articles critical of the project came out in March 2014. They described the project as "an Internet portal for the homeless". 14 It was argued that the government had spent a huge amount of money on a useless IT solution. Many of these reports, unfortunately, duplicated the harmful stereotypes about the poor and their digital exclusion. Speaking out in the media, even social workers indicated that the requests submitted via the internet might be treated as suspicious, because poor people are assumed not to have access to new technologies.
Implementation of Emp@thy also caused a reaction from public institutions. The Supreme Audit Office ( Najwyższa Izba Kontroli, NIK) noted a number of irregularities during the development of the projects’ functions (related to the management of public funds and the fulfilment of the goals). 15 The project was also under investigation by the Central Anti-Corruption Bureau. 16
In addition to these problems, various objections strictly related to human rights appeared. From the perspective of the right to social assistance, digitalisation was developed to ensure citizens better information on available services, improve contact with the administration, and increase accessibility of benefits. Unfortunately, the project hardly meets all these objectives.
The Emp@thy portal provides information about welfare benefits in a very unapproachable and bureaucratic manner. Very often it simply duplicates legislative language. Unfortunately, the portal does not allow people to ask questions and contact the administration directly. The data also show limited interest among portal users. 17 As of November 2015 there were over 468,000 visits to the portal, but in 47% of the cases, users only went to the main page and did not use other functionalities. According to research, 70% of people receiving help through local social assistance offices know about the Emp@thy portal, but 62% of them could not indicate what sort of information or services they could receive through this tool.18
Other problems are related to the e‑form, which allows the submission of a request for assistance through the internet. There are a number of restrictions in this area, related mostly to the authentication procedure. To submit an application this way, citizens need a specific e‑profile on the ePUAP platform. However, only a small percentage of Poles use any of these authentication tools. The portal also only allows the submission of limited types of applications, which do not cover the whole spectrum of the social policy area. As a result, citizens submitted a very small number of applications through the portal (in 2014-2016 only 683). 19
Moreover, the audit by NIK showed that officials do not have the appropriate skills and knowledge about the new IT system. One of the most disturbing examples of this situation was that some local social assistance offices had not opened or did not even know about the special mailbox where e‑forms from citizens would appear. 20
On the other hand, there is a potential risk of digital exclusion while developing e‑government services, especially when the administration starts to rely more on e‑forms than direct contact with the officials. This is not yet an issue in Poland. However, we could easily imagine that it could be one, affecting those groups who do not have access to the internet or do not have the skills to use the internet. In Poland these are vulnerable groups such as elderly people, people from poorer regions or less educated people. 21
To some extent, one of the Emp@thy project’s successes might be the online verification of citizens’ data from other offices. This mechanism makes the application procedure easier for citizens, and allows them to save some time (e.g. they are not obliged to show various certificates from other offices). Inspections also show that only 16% of local offices still do not use this possibility, so it has become rather popular. 22 However, in this area objections may appear related to the right to privacy. Concerns regarding the processing of personal data have arisen over the Emp@thy project. 23 These were mainly due to the creation of new public databases that centralised a lot of personal data – including sensitive data. The new central registers contain information not only about the people receiving benefits, but also about members of their families (more than four million people). The retention period was set for a very long time – ten years after the person stopped receiving benefits.
Other reservations were related to the rather lax rules for the exchange of information on citizens between social assistance and other authorities (e.g. tax offices). While such a mechanism is beneficial to citizens, the exchange rules should be precise and guarantee the confidentiality and safety of personal data. But what is even more disturbing is that NIK’s inspection demonstrated that for more than one year the databases created under the Emp@thy project functioned without any legal grounds. 24 . The right to data protection and the right to privacy can only be limited under specific conditions. The lack of legal grounds for the public administration to process private data is a clear violation of these standards.
Another issue is the welfare fraud tracking component. In accordance with international standards of protection of the right to social assistance, rules for granting and limiting social assistance should be as transparent as possible. The automatic detection of fraud may undermine this principle. Official documentation explains that welfare fraud detection is primarily used to "search for non-compliance". 25 But what non-compliance means was not described precisely. When the system generates information about inconsistencies, the official who is responsible for the case should clarify the matter. A problem may, however, arise when on the basis of an inconsistency being detected, the benefit is withdrawn without precise explanation. When so little is known about how the system works, the possibility to challenge the information about non-compliance can be difficult. However, to better assess the risks associated with the automatic fraud-detection system, more information should be collected on the practical use of this mechanism. Inspections conducted by NIK only show that this mechanism was hardly used to detect any kind of fraud (0.7% of all local social assistance offices indicate the practical usefulness of this tool). 26 So there may be legitimate doubts as to whether this mechanism is necessary in the daily work of social assistance offices.
Profiling the unemployed
In May 2014, the Polish government introduced a reform of the functioning of local labour offices (these are public entities responsible for providing benefits and active labour market programmes for the unemployed). 27 One of its significant elements was to implement profiling of assistance for the unemployed. This was an example of the new generation of services which lead to greater individualisation through the use of new IT solutions. The MRPiPS also specified that the main reasons for the introduction of profiling were to counteract unemployment more effectively, increase the efficiency of labour offices and guarantee public services of a higher quality.
In practice, profiling involves dividing all unemployed people into three categories, taking into account their individual characteristics. Assignment to a given category determines the type of assistance that a particular person can receive (e.g. job placement, vocational training, apprenticeship, activation allowance). Categorisation is based on data collected during a computer-based interview with the unemployed person, which follows a guide prepared by the MRPiPS for its employees. A total of 24 different dimensions (like gender, age, education, disability, degree of disability) are reported in the electronic database and each of them is assigned a score (0 to 8). An algorithm determines the result of the profiling process. Labour office employees can change the results of the categorisation in exceptional circumstances. Unemployed people can only request to have their profile done again if their life situation has changed since the profile was established. 28
During the legislative process regarding the reform of labour offices and throughout the implementation process, profiling was criticised from a human rights perspective. The Inspector General for Personal Data Protection ( Generalny Inspektor Ochrony Danych Osobowych, GIODO)29 and the Ombudsman30 filed reservations as to the compliance of the new solutions with the Polish constitution. It has been stated that the regulations on profiling are not precise enough. Also, there have been claims as to the lack of adequate guarantees for the protection of the right of privacy and personal data (i.e. there being no transparent, legally regulated procedure enabling a change of the assigned profile). Critical arguments on profiling have also been presented by Panoptykon Foundation. 31 The organisation raised arguments related to the protection of privacy, the non-discrimination principle and the right to work and social protection.
Research conducted in 2015 (after new provisions came into force) has shown many problems related to the profiling mechanism. 32 First, it was proven that there is a systematic problem with transparency. The existing law fails to specify what the determination of categories looks like. The criteria according to which a certain profile of assistance is attributed to a person remain unknown to the unemployed throughout the whole process of profiling. They remain unclear even to the staff involved in this process. The unemployed are also deprived of the right to obtain information about the logic behind profiling; in particular they cannot verify how certain features affected the profile of assistance that was attributed to them. Lack of transparency in this area violates international standards of providing public services.
Another problem is related to the principle of equal treatment. Allocation to a given profile is determined on the basis of features such as age, gender or disability. In practice it may be the case that the situation of specific unemployed individuals is differentiated on the basis of the criteria listed above, which may be considered discriminatory practice. In the context of the ICESCR this would mean a breach of one of the minimum core obligations related to the right to work. The research concludes that the Polish system may actually result in limiting access to specified active labour market programmes for disadvantaged groups among the unemployed. In such cases, profiling does not fulfil the role of a measure of affirmative action accounting for specific needs of people who have difficulties finding employment.
Equally problematic is the automatisation of the decision-making process in the profiling mechanism. The creators of the new system proceeded from the assumption that a computer will work better than a human and that decisions taken in an automated manner will be more objective. However, the mechanism used in Polish labour offices is based on simplifications that do not take into account the very wide variety of complex life situations. For example, the range of causes of unemployment does not include homelessness or a stay in prison, which keep many people out of employment. The rules that regulate profiling are very rigid and do not allow for an unemployed person to actively participate in the process. For instance, the law does not provide a possibility to demand that the profile be changed (with the exception being if their life situation has changed after profiling) or re-verified in the case when the unemployed person him/herself thinks that they qualify for another profile or that an error has been made in the course of its determination.
According to official statements, each profile was described as equally good in terms of the assistance received. However, the way that the profiling was designed shows that people assigned to profile II can get a much wider range of support than those that fall under profile I, and especially those that fall under profile III. 33 For example, people assigned to the third category can take part in special programmes. But in practice, they are very difficult to organise, which is confirmed by statistics: 38% of local labour offices do not offer any programme addressed to the third profile.
The documentation also shows that labour office staff are instructed to treat unemployed people according to the principle of "limited trust". This applies particularly to people qualifying for profile III. They are officially described as lacking motivation and having a “calm” nature. Documents indicate to staff that the unemployed can manipulate and deceive officials just to be assigned to the desired category. This logic of "suspicion" translates into, for example, a lack of legal guarantees to access information on the processed data and the results generated by the computer.
According to the Polish constitution and international law, human rights standards should play a crucial role in creating and developing public policies. However, digitalisation and implementation of e‑government in the welfare sector show that the prospect of fundamental rights is virtually unnoticeable or plays a minor role. Analysis of strategic documents and laws on implementing new tools shows that these documents focus primarily on financial, technical and administrative issues. Somewhere in the background there is the question of access to specific forms of support (which may be a reference to the general standards of social rights). Each new piece of legislation which introduced new public IT systems was also assessed from a data protection perspective by GIODO. However, the recommendations of this institution are rarely taken into account in the legislative process.
The examples that we described above show that widespread digitalisation in public administration could behave like a “double-edged sword” for human rights. IT tools could help in better realising the right to social protection or the right to work. Well-designed web pages with transparent and clear information about certain welfare benefits could help people to gather knowledge about possible assistance. Submitting applications for benefits via the internet saves time and reduces what is sometimes unpleasant contact with the authorities. These measures have a potential empowerment impact – they could make it easier to navigate the bureaucratic system, and in the end could positively affect the availability of social services. Unfortunately, in practice these good examples are reduced by the problems in developing, managing and maintaining the IT infrastructure.
On the other hand, digitalisation could limit certain human rights. Firstly, when public administration emphasises developing internet-based tools, this could lead to digital exclusion and discrimination. Secondly, there is a risk related to the automatisation of the decision-making process (detecting welfare fraud or using profiling techniques for assigning support). This trend can lead to a decision-making process that is discriminatory and not transparent and which does not meet international human rights standards. Thirdly, digitalisation of certain public services can lead to a violation of the right to privacy and data protection. Creating new big databases and ways to process and analyse information should always be consistent with the principles of necessity and proportionality. As we mentioned above, concerns related to data protection and safety rules are one of the biggest barriers for Poles to use more e‑services. Reliable and clear data protection safeguards could definitely build more trust between citizens and authorities and encourage the use of more e‑government products and services. The link between privacy and social services is a good example of how civil and political rights and ESCRs are indivisible, interdependent and interrelated.
Based on our observations, we recommend the following steps:
At the national level:
The creation and implementation of a mechanism to evaluate public IT systems from the point of view of human rights. This mechanism should include both political and personal rights, as well as economic, social and cultural rights. Policy and legislation, as well as IT systems themselves – such as the use of algorithms – should be subject to this assessment.
It is necessary to develop appropriate safeguards for the protection of individual rights in the context of information systems which support public services. The protection of personal data is not enough. In the era of automated decision making based on the processing of data we need a new approach to principles of transparency, due process and the place of the citizen in the whole process. Such a discussion could be carried out with the participation of public institutions, academia and civil society organisations.
At the international level
The treaty bodies dealing with the protection of economic, social and cultural rights should be interested in the relationship between these rights and the internet and new technologies. In particular, the Committee on Economic, Social and Cultural Rights should conduct a day of general discussion on this subject, and at the end of this elaborate on the issue in a general comment. This document should take into account both the benefits and risks arising from the use of new technologies in social policies.
1 Śledziewska, K., Levai, A., & Zięba, D. (2016). Use of E-Government in Poland in Comparison to Other European Union Member States. Information Systems in Management, 5(1), 120.
2 Ibid., 121.
3 In this case understood as submitting completed forms using websites of public authorities.
4 Eurostat. (2015). Reasons for not submitting completed forms to public authorities' websites. www.ec.europa.eu/eurostat/web/products-datasets/-/isoc_ciegi_rtx
5 Popiołek, M. (2013). E-government in Poland – selected issues. Journal of Education Culture and Society, 2, 402; Ministry of Administration and Digitization. (2012). Państwo 2.0. Nowy standard dla e-administracji, 63-64.
6 The Polish Committee of the European Anti-Poverty Network. (2016). Position concerning the implementation of ICESCR in Poland. www.tbinternet.ohchr.org/Treaties/CESCR/Shared%20Documents/POL/INT_CESCR_ICO_POL_23032_E.doc
7 For example, Act on Personal Data Protection (1997), Act on the Promotion of Employment (2004), Anti-Discrimination Act (2010).
8 Committee on Economic, Social and Cultural Rights. (2008). General Comment No. 19: The right to social security (art. 9), E/C.12/GC/19, para. 10-27. tbinternet.ohchr.org/_layouts/treatybodyexternal/Download.aspx?symbolno=E%2fC.12%2f%20GC%2f19&Lang=en
9 Ibid., para. 59; Committee on Economic, Social and Cultural Rights. (2006). General comment No. 18: The right to work, E/C.12/GC/18, para. 31. tbinternet.ohchr.org/_layouts/treatybodyexternal/Download.aspx?symbolno=E%2fC.12%2fGC%2f18&Lang=en
10 Committee on Economic, Social and Cultural Rights. (2008). Op. cit., para. 24.
12 MRPiPS. (2014). Brochure about Empathy project – Czym jest empatia.
13 The Electronic Platform for Public Administration Services (ePUAP) was designed and developed to allow public institutions to make their electronic services available to the public. More about the platform is available here: www.epuap.gov.pl
14 See, for example, Zieliński, R. (2014, 25 March). Powstał portal dla osób bezdomnych. Skorzystają te, które mają laptopy, Dziennik Gazeta Prawna. www.gazetaprawna.pl/artykuly/786190,portal-dla-osob-bezdomnych-z-laptopami.html
16 Skłodowski, T. (2015, 11 May). "Empatia" pod lupą CBA i prokuratury. Nieprawidłowości przy przetargu na portal Ministerstwa Pracy? Polska Times. www.polskatimes.pl/artykul/3857455,empatia-pod-lupa-cba-i-prokuratury-nieprawidlowosci-przy-przetargu-na-portal-ministerstwa-pracy,id,t.html
17 NIK. (2016). Op. cit., 73.
18 Ibid., 76.
19 Ibid., 74.
20 Ibid., 76-77.
21 Ministry of Infrastructure and Development. (2014). Digital Poland for 2014-2020, 10. www.polskacyfrowa.gov.pl/media/10410/POPC_eng_1632015.pdf
22 NIK. (2016). Op. cit., 68.
23 They were raised by the Inspector General of Data Protection and the Panoptykon Foundation. For more information see Panoptykon Foundation. (2014, 13 October). Pomoc (czy policja) społeczna? https://panoptykon.org/wiadomosc/pomoc-czy-policja-spoleczna
24 NIK. (2016). Op. cit., 54.
25 MRPiPS. (2014). Powykonawcza dokumentacja podsystemu CBB – architektura logiczna, architektura techniczna oraz wymagania funkcjonalne i niefunkcjonalne.
26 NIK. (2016). Op. cit., 84.
27 MRPiPS. (2013). Uzasadnienie projektu ustawy o zmianie ustawy o promocji zatrudniania i instrumentach rynku pracy. www.sejm.gov.pl/sejm7.nsf/PrzebiegProc.xsp?nr=1949
28 MRPiPS. (2014). Profilowanie pomocy dla osób bezrobotnych. Podręcznik dla pracowników powiatowych urzędów pracy. www.panoptykon.org/sites/default/files/podrecznik_profilowania.pdf
29 GIODO. (2013). Uwagi dotyczące projektu nowelizacji ustawy o promocji zatrudnienia i instrumentach rynku pracy z dnia 23 października 2013. www.legislacja.rcl.gov.pl/docs//2/171820/171829/171833/dokument89898.pdf
30 The Ombudsman. (2014). Wystąpienie Rzecznika Praw Obywatelskich w sprawie profilowania osób bezrobotnych z dnia 19 listopada 2014.
31 Panoptykon Foundation. (2014). Stanowisko dotyczące projektu rozporządzenia o profilowaniu pomocy dla bezrobotnego z dnia 15 stycznia 2014 roku. www.panoptykon.org/sites/default/files/panoptykon_profilowanie_bezrobotnych_rozporzadze nie_opinia_15.01.14.pdf
32 Niklas, J., Sztandar-Sztanderska, K., & Szymilewicz, K. (2015). Profiling the Unemployed in Poland: Social and Political Implications of Algorithmic Decision Making. Warsaw: Panoptykon Foundation. www.panoptykon.org/sites/default/files/leadimage-biblioteka/panoptykon_profiling_report_final.pdf
33 Profile I covers mainly active, young and mobile persons, having appropriate professional qualifications and interpersonal skills. Profile II typically includes persons who have certain professional skills, but unfortunately are redundant in the labour market, or worked for a very long time in one company. Profile III comprises persons with serious life problems or those who do not want to cooperate with the employment office. The guidelines suggest that the following people should be placed under this category: persons with disabilities, single women raising children, persons registering themselves only because of the need to obtain health insurance, or persons from small towns who are isolated from larger urban areas.