Limited knowledge and informal lobbying: internet regulation through content filters in Swedish public libraries

Abstract

Purpose

The purpose of this paper is to describe and explore the current state of internet regulation through content filters in Swedish public libraries.

Design/methodology/approach

Data was collected through an electronic survey directed to library managers of Sweden’s 290 main municipal libraries. 164 answers were returned, yielding a 57% response rate. The analysis comprises descriptive statistics for quantitative data and an activity theory approach with focus on contradictions for qualitative counterparts.

Findings

In total, 33% of the responding libraries report having content filters; 50% have not; and a surprising 18% do not know. There is a strong correlation between internet misuse and positive attitudes towards filters, and, reversely, between lack of misuse and lack of active stances concerning filters. Rather than seeing this as weakness, the authors suggest that there is strength in a context-bound flexibility open to practical experience and weighting of values, ethics, legislation and local circumstances. More troublesome indications concern the high deferral of decision-making to local authorities (municipalities) whereby libraries are left with limited insight and influence.

Research limitations/implications

The situation calls for professional organisations to address political mandate questions, and educational programs to strengthen future information professionals’ knowledge of IT in general; filter issues in specific; and local authority decision-making. The study highlights the need of adequate information professional competences and mandates to decide on and oversee internet regulation.

Originality/value

To the best of the authors’ knowledge, this is the first internationally published study on content filters in Swedish public libraries.

Citation

Johansson, V. and Lindh, M. (2023), "Limited knowledge and informal lobbying: internet regulation through content filters in Swedish public libraries", Journal of Information, Communication and Ethics in Society, Vol. 21 No. 3, pp. 243-258. https://doi.org/10.1108/JICES-12-2022-0105

Introduction

This article addresses a lack of empirical knowledge concerning public libraries’ conditions for and contradictions in internet provision and regulation by presenting results from a national Swedish study. All internet provision in public libraries is regulated in some ways: through automated content filters at internet suppliers, in local networks or individual computers; or through social and manual control measures such as visual monitoring, identification requests, review of search logs and open placement of computers. These regulations are intimately connected to ethics, human rights principles, national and international jurisdiction and local policies (Buchanan, 2008; Frické et al., 2000; Knox, 2011). Here, ideas on human rights of freedom of information cross paths with, and need to be balanced against, concerns for privacy and ownership, and unethical and illegal content and services such as child sexual abuse material (CSAM). Consequently, issues of internet provision in public libraries also engage numerous actors with strong interests and positions vis-à-vis the library, including local authorities, schools and parental interest groups, which creates many challenges for the public libraries (Brown and McMenemy, 2013; Kann-Christensen and Pors, 2004; Muir et al., 2016; Pors, 2001).

To guide libraries in this conflicted terrain, professional library organisations have published numerous position statements, declarations, guidelines and codes of ethics (henceforth normative documents). Among these, the American Library Bill of Rights represents a particularly strong stance, admonishing libraries to protect and ensure unreserved freedom of information to all users regardless of “origin, age, background or views” (ALA, 1939/1996). In a similar vein, the international Glasgow Declaration advocates “unlimited access to information” and states that libraries should “oppose every form of censorship” [International Federation of Library Associations and Institutions/Freedom of Access to Information and Intellectual Freedom (IFLA/FAIFE), 2002]. The Alexandria Manifesto states that obstacles to free provision of information are to be amended whether of “structural or other art” (IFLA, 2005) and the Internet Manifesto that “obstacles to the information flow [on the internet] should be removed” (IFLA, 2014). Likewise, the IFLA Code of Ethics rejects “the denial and restriction of access to information and ideas” (IFLA, 2012). Strong correlations between intellectual freedom and democracy permeate all these documents.

Despite the centrality and weight of the above-mentioned normative documents within their respective domains, they are nevertheless subject to critique. Frické et al. (2000) argue from an analysis of ethics theories that the ALA’s demands for unlimited access to information is based on a misunderstanding and that limitations to information – including censorship – constitute a legitimate measure in certain cases to protect even more fundamental human rights. From a more consequentialist perspective, Kann-Christensen and Pors (2004) alongside Muir et al. (2016) emphasise that evading automated regulation measures in the form of filters risks an increase of alternative regulatory strategies that are of a more privacy invasive character. Other researchers point out the more general troubling lack of knowledge concerning to what extent and how normative documents feature in and shape actual regulation activities in local public libraries (Ferguson et al., 2016; Luo, 2014; Rubin and Froehlich, 2011).

In response to some of the knowledge gaps outlined above, this article draws on results from the larger research project Internet Use in Public Libraries – Policy, Practice, and Pedagogy (Johansson and Lindh, 2020) including a national survey directed to Swedish public libraries (henceforth libraries). Through further analysis of the project’s extensive quantitative and qualitative survey data, the aim here is to describe and explore the current state of internet regulation through content filters in Swedish libraries. The analysis comprises descriptive statistics for quantitative data and an activity theory approach with focus on contradictions for the qualitative counterparts. Three questions guide the analysis:

RQ1.

To what extent and under what conditions is internet provision and use regulated through content filters in Swedish public libraries?

RQ2.

What contradictions can be identified in relation to filter regulation issues in the libraries?

RQ3.

How do the libraries act on these contradictions and in what ways do normative documents figure in such activities?

Literature review

Studies on content filter use in libraries point to considerable national differences, yet also illustrate a common tendency of increasing usage over time. Two Danish studies from the early 2000s show that filtering was rare at the time yet doubled over the short study period from 3% in 2001 to 6% in 2003 (Kann-Christensen and Pors, 2004). In Australia a decade later, 44% of libraries reported using filters on their public access computers. Most filters were set to block out the broad and fuzzy category of “offensive content”, but also “very large files”, file sharing, games and social networking sites (ALIA, 2013). About 40% had dedicated terminals for children, and 35% of these did not provide links to third-party material. The same year, Brown and McMenemy (2013) found that all 31 respondents from Scotland’s 32 local authorities used content filters in their public library services. Some years later in the UK, corresponding figures show that 100% of respondents filtered internet access on library computers, and 83.6% filtered Wi-Fi access (Muir et al., 2016). Filtered content was described as sexual, hacking, violent, intolerant/hate and extremist. Similar categories feature in Brown and McMenemy’s (2013) research, but with greater emphasis on illegal content, and the addition of gambling/gaming and social media sites. In the UK study (Muir et al., 2016), higher levels of filtering for children were the norm, sometimes combined with “walled garden” solutions (only pre-approved material) on children’s computers.

Some of the studies also investigated library staff and management’s motives for and attitudes towards filters. The Scottish study shows that main justifications for filtering software concerned prevention of access to illegal and/or inappropriate material, followed by a desire to protect children and vulnerable users from inappropriate material; considerations of duty of care and reputation; alongside protection of the computer network (Brown and McMenemy, 2013). In the UK study (Muir et al., 2016), although most of the respondents said that serious internet misuse incidents were rare, the general attitude was that filters were a practical solution to a problem. Safeguarding children in particular was a clear justification, and moral pressures from parents and teachers were also mentioned. However, filtering adults’ internet access also seemed to be accepted with “at most, mild regret” (Muir et al., 2016, p. 97). This is also the only study to include user attitudes to filters; 89% of the libraries made users aware of the filtering, and about 66% had received filter related complaints from the public – usually concerning incidents of over-blocking and rarely about the existence of filters as such. Overall, the study describes a general acceptance of the need for filtering among both staff and users.

Mandates to decide on regulation issues in the library is also a vital point of concern, particularly as these decisions are often taken by authorities outside of the libraries. In the Scottish study (Brown and McMenemy, 2013), the source of the filter policy decision (whether for or against) was in 56% extended (non-library) senior management/local authorities; 22% library management; 16% joint library-local authority; and in 6% unknown/undisclosed. Moreover, only 3% stated that frontline library staff had the ability to immediately release blocked content if deemed appropriate, whereas 88% stated that content could not be released at point of use but had to be considered and released later if deemed appropriate. In total, 6% had no such procedure in place at all. In Muir et al.’s (2016) UK study, decisions of filtering software were also most often made at the corporate level (48.8%) rather than by library managers (30%). Some also indicated that the filtering itself was implemented at a corporate level, meaning that the same types and levels of filtering used in, e.g. schools were also applied in the libraries. Less than 10% allowed front-line library staff the mandate and responsibility to respond to unblocking requests, more commonly deferring such response and decisions to library service senior management (52.5%), IT staff at local authority level (41.3%) and IT staff at library service level (41.3%) – alone, or in combination.

Overall, Muir et al. (2016) emphasise that filters are a challenge to intellectual freedom. Libraries should avoid filters and library professionals should take active part in public debate and educate the public on the values of freedom of information and privacy, which are at risk of being dominated by child protection and anti-pornography agendas. Brown and McMenemy (2013) also note the risk, given the spread of categories being blocked, that “objectionable material” will be conflated with illegal material, of which they describe only the latter to constitute a valid category for libraries to block with the aid of filters. There is agreement among several of the studies on the importance of user internet education and stricter controls on child access (Brown and McMenemy, 2013; Muir et al., 2016). Brown and McMenemy (2013) also emphasise the values of alternative regulatory strategies such as acceptable use polices (AUPs), privacy screens and recessed monitors. Their harshest critique concerns the libraries’ lack of knowledge and lack of power to decide on and manage regulation issues on their own, and they find staff knowledge on information and communication technology and filters in general and local filter policies in particular, the use of AUPs, and the existence of clear unblocking procedures “unacceptably” low or altogether lacking in the studied libraries.

How libraries actually position themselves and act on internet regulation issues in practice seems even more rarely studied. However, existing studies indicate that normative documents are relatively rarely used, in favour of other sources, primarily colleagues (Luo, 2014). The most common identified ethical conflict among information professionals in a study covering Britain, Ireland and Australia was between professional ethics and organisational requirements, and problems were found to be referred “[…] ‘upwards’ through the organisational hierarchy rather than dealt with using library codes of ethics” (Ferguson et al., 2016, p. 548).

Theory

For the qualitative analysis of contradictions, activities and normative documents (RQ2 and RQ3), Engeström’s (1987) “third generation” of activity theory is invoked. The unit of analysis in this perspective comprises the purposive activities through which subjects (here: libraries and municipal authorities/IT departments) act on a shared object (internet regulation in libraries) with a transformative intent. Elementary school politicians, parental organisations and library users constitute elements of the wider context, described in activity theory as community/actors. Tools denote whatever is used by the subjects to accomplish the object, and we are especially interested here in how normative documents may feature as such. Previous research also suggests that division of labour between libraries and municipal authorities/IT departments, describing how work is divided with consequences for roles and hierarchies within and between activity systems, and rules (such as legislation and norms) also constitute interesting aspects of the analysis. Taken together, these elements form an activity system, within which motives for action arise through unfolding contradictions between elements of the system. This is also what we guide our analysis towards.

Method

This article provides an extended analysis of data from the larger Swedish research project Internet Use in Public Libraries – Policy, Practice, and Pedagogy (Johansson and Lindh, 2020; c.f. also Acknowledgements). This overall project includes interviews and document studies, but we focus here solely on the survey data, as these other empirical data sources primarily confirmed the more extensive quantitative and qualitative filter related findings of the survey.

Data collection

The total number of integrated and freestanding public library service points in Sweden at the time of study was 1,091 (National Library of Sweden, 2019). Of this total population, the survey was directed to the main public library in each of Sweden’s municipalities, providing a sample of 290 libraries. The questionnaire was distributed in electronic form and open for 12 weeks between December 2019 and February 2020. Reminders to non-respondents were sent out five times during this period. Of the 290 libraries, 164 answered the survey, yielding a response rate of approximately 57%.

The survey questions were formulated based on an extensive review of previous research and consisted of a combination of multiple choice and free text questions. However, even though the high level of autonomy concerning library related decisions for the Swedish municipalities is likely to have resulted in different types or brands of filters and differences in filter level settings (where used) between libraries in different municipalities, our survey did not ask for details about this in order to minimise risks of researcher and participant bias. For similar reasons, the questions did not explicitly ask for opinions and uses of normative documents. The survey form is available for consultation and reuse under a CCBY 4.0 license as a stand-alone publication in English translation (Johansson and Lindh, 2022).

Data analysis

For the analysis answering to RQ1 concerning extents and conditions for filter use, basic descriptive statistics was used. This amounts to retrieval and provision of raw numbers and percentages for fixed answer questions relating to, e.g. use of filters or not; who is responsible for filter decisions; and libraries’ attitudes to the aforementioned. Due to the relatively small and non-representative sample, the percentages reported in the results section have been rounded off to the nearest whole figure throughout.

For the qualitative analysis answering to RQ2 and RQ3, activity theory informed a directed content analysis approach focused on identifying themes of contradictions between the central elements of activity and of the role of normative documents as tools in related library activities (Hsieh and Shannon, 2005).

Research ethics

The study adheres to international social science research ethics and principles (Bryman, 2016). Participation was informed, voluntary and anonymous. Special care was taken to remove any accidental identifying traits from free text answers prior to analysis and reporting.

Limitations

Concerning quantitative limitations, the 57% response rate for the sample population of main libraries in each Swedish municipality yields a confidence level of about 95% with a 5% error margin for this sample. Calculated in relation to the total population of Swedish public library service points, the confidence level drops to about 90% with a 6% error margin, which although weaker still provides ground for strong indications. Considering limitations of a more qualitative character, other due caution is warranted. The survey was directed to library managers and/or IT librarians and similarly positioned staff, who were asked to answer on behalf of their library in a more general sense. Possibilities to distinguish and explore the views and knowledge of individual library staff within the participating libraries are therefore lacking from this data. The study also lacks first person perspectives of municipal boards/politicians and their IT departments, as well as of library users.

Results

Extent of and conditions for content filters

One-third (n = 54, or 33%) of the 164 libraries in the study stated that some form of content filter was used on all or some of their public computers, whereas half of the respondents (n = 81, 50%) answered no. Somewhat surprisingly, 18% of respondents answered that they did not know. Of the 54 libraries with filters, only six use those filters on only some of their public computers. And of those six, only one stated that the reason is to restrict internet access for under aged library users (in this case, under the age of 15). The reasons for partitioned filter use in the remaining five libraries are not revealed through the survey. Respondents from the 54 libraries with filters further describe that these filters are intended to block a relatively broad spectrum of content, with pornographic material – both illegal (CSAM and similar) and legal – most often mentioned. Medium frequency occurrence comprises content relating to racism/incitement against ethnic/cultural groups, violence, terrorism, radicalisation, illegal drugs and different types of games and gambling services. But as previously indicated, there is a high degree of uncertainty here as well: almost 40% of respondents from libraries with filters are unaware of what those filters are intended to block.

Among the third of respondents (n = 54; 33%) who were certain that they have filters installed, this decision has predominantly been made at the municipal level without consulting the library (70%). Another 20% of these libraries had arrived at a joint decision with their municipalities, and in only 7% of cases was this decision the library’s own. Among the subset of respondents who were certain that they do not use filters (n = 81, 50%), this decision had been made by the library itself in 21% of cases; together with the municipality/IT department in 9%; and in about 8% by the municipality/IT department alone. Interestingly, in the majority of cases (48%), this indicates that the lack of filters is the result of inaction – rather than an active decision – on behalf of all subjects. Respondents in both groups (i.e. with and without filters) further report little internal or external disagreement on the current filter state of their library, although with slightly higher occurrence in libraries with filter (13%) than without (9%).

The responses also reveal that information to users about, and user input on, filter issues is rare. Most responding libraries with filters (63%) never receive comments or complaints concerning their filters. The remainder stated that it seldom (33%) or sometimes (4%) happens. However, the extent to which the existence of content filters is known to the users should reasonably affect possible user feedback on the issue, and the results show that such information is sparse. Of the 54 libraries that were certain that they have filters, only 12 (22%) stated that they offer their users information about this. For this information, a variety of measures are used: when the user registers for a loan card (four respondents); through pop-ups if a search or search result is blocked (four); on notes near the computers (three); on the screen when a user logs on (two); or somewhere else in the library facilities (one). Nevertheless, 78% of respondents with filters provided no information to users about their existence. Of the libraries with filters, 28% expressed the opinion that they provided enough information to the users about their filters; 30% thought that they do not; and the majority, 43%, marked uncertainty or lack of an opinion.

Motives

Turning now in this qualitative part of the results presentation to activity theory, contradictions appearing in relations between subject(s) (here, libraries, municipal boards and their IT departments) and object (internet regulation issues) should point to motives regarding perceptions and actions for and against filters. Such motives appearing in our analysis are found to both favour and oppose filters, from technical as well as social perspectives.

Active and deliberate stances against filters display a range of grounds. Some arguments concern how filters can provide a false sense of security and may under block or easily be circumvented. Another argument is that filters are blunt tools and may over-block. Furthermore, respondents at libraries with filters (n = 54) described motives for filter use as partly overlapping with wishes to prevent certain types of content and activities, and to protect certain values and individuals. The most common motives for filter use are intentions to prevent content and activities that are clearly illegal (e.g. CSAM and illegal filesharing) and technically harmful (e.g. viruses and malware that can damage software and users’ information and integrity). Perceptions in this category are clear: “[It’s] part of the municipality’s policy. [And I] don’t think that tax funded technology should facilitate opportunities to commit crimes either”. But there is also often a clear reactive element in arguments for filters as several respondents reported that the filters were installed as reactions to previous problems with internet misuse.

Other content and activities blocked by filters are described as inappropriate in other ways, like searching for and looking at pornographic material of non-illegal character. Because of its “fuzzy” nature, abuse is another interesting concept that frequently recurs. By blocking “inappropriate” content and activities, library users are seen to be protected from harm. Filter use for such purposes is thereby described as relatively uncomplicated.

Additional motives of protection are to safeguard users, the shared library environment and the library’s central values in a broader sense through the use of filters. The most frequent argument here concerns the desire to ensure, through technical blocking, that children are not at risk of being exposed to inappropriate material through others or their own inappropriate or inconsiderate internet use. However, our data show that very few libraries only apply filters to dedicated children’s computers (c.f. above). Rather, associated protections are often extended to all library users regardless of age.

Another common argument is that filters make it easier to protect and uphold a shared environment where everyone feels comfortable, and not violated, scared, or shut out in any other way through the information activities of others in the room. Related to this, we also see that concrete experiences of problems with misuse of the internet in libraries are often connected to a positive attitude towards filters: “Seems reasonable according to the employees. We’ve all come into contact with ‘deviants’ when both computers and printers have been used for criminal or half-criminal activities.” Such experiences are also often connected to libraries actively requesting more and better content filters.

Among respondent libraries without filters (n = 81), the arguments are also varied, ranging from more or less neutral approaches to arguments based on clear statements against filters. Neutral statements express that even though filters are not used, they are not opposed to filters if they should be needed – provided they were to function satisfactorily: “Not [an] active [stance] really, it’s mostly that […] we haven’t thought that there’s anything that works well enough.” The absence of filters as well as of opinions for or against their use seems to be related to contexts in which the question has not been deemed important or considered at all. In this group of respondents, filters are often described as a ‘non-issue’, as something that has not presented itself as either a practical problem or an ideological/political issue and thereby has not been actively discussed: “this [absence of filters] hasn’t been our decision [but] [n]o schism has come out of that. We don’t have problems with our users’ uses of the public computers.” Thus, respondents without filters largely expressed the neutral view that filter absence is relatively unproblematic, but also that a potential installation of filters would be equally unproblematic, if a practical need arises.

The widespread absence of stances in either direction for or against filters is echoed by an apparent lack of negative response on behalf of users. Survey comments primarily mention user questions raised in relation to games and gambling in various ways. Sometimes children/youth try unsuccessfully to download or access certain games; sometimes adults try to access gambling sites and enquire about why they cannot do so. Misuse of internet is described as rare or virtually non-existent. Inappropriate uses most frequently mentioned concern pornography, violent material, betting/gambling and racist material. Reasons given for such inappropriate uses include ignorance due to users not understanding the AUPs at login; mischief by users of school age; and perceived mental illness. No comments mention incidents of a serious kind. Overall, the qualitative responses support and further explain how absence of filters is mainly a question of absent decision or stance concerning filters altogether. Respondents appear unmotivated to take a stance on filters whilst users’ Internet habits do not cause significant problems.

Division of labour

Another theme of contradictions emerging in the analysis concerns division of labour for internet regulation in the libraries, which leads to issues of hierarchical distributions of knowledge, competences and influence over decisions. Respondents explicitly associated the libraries’ overall low knowledge about the existence of and settings/functionalities for filters (c.f. above) with the fact that municipalities and their IT departments have the responsibility to make these decisions: “Our settings are managed by [the] IT [department] of [X] city since we are part of the [Y] region’s libraries, so I can’t answer that question.” Nevertheless, only a few respondents reported being shut out from discussions and decisions that they would like to be a part of concerning filters; this appears to be a rare situation. More often, our data suggest that the municipalities and their IT departments take the lead because they have developed pre-existing active stances regarding content filters (n.b. both for and against) whilst most libraries have not seen filters as an urgent issue – neither to avoid nor request.

However unsatisfactory, the libraries generally had considerably greater knowledge about the existence of filters than how they are designed – i.e. what is blocked, how and why: “The municipality has a filter on all computers, but we don’t know exactly what it blocks. But they also intend to block things that can hurt the computers.” Respondents consistently explained this common lack of knowledge by referring to the fact that it is the municipality’s and/or its IT department’s responsibility to decide on and manage issues relating to internet, information safety and filters in public libraries, by virtue of the library’s placement within a municipal network. The theme reveals a far-reaching division of labour concerning internet filters that not only results in lack of knowledge in libraries, but also to a widespread and seemingly welcomed or at least non-problematised deferral of responsibility for the overall issue from the libraries themselves to their municipal authorities.

A similar unequal division of labour with potential negative consequences can be discerned in relation to users as well. Overall, the libraries do not seem to pay much attention to the question of user information about the existence and design of content filters. As with filters, this situation often appears to be the result of inaction or lack of decision. Only one respondent explicitly expressed that filter information is unnecessary: “Why should we inform about that.” Other respondents deemed that their filter information works well and is satisfactory, even when users must proactively ask for it themselves: “There are plenty of opportunities to ask at the desk if they have any questions.” In similar ways, information that is only provided as the result of a blocked information search was described as fully satisfactory.

Rules and community/actors

As a third qualitative analytical theme, the elements rules and community/actors merge in our analysis into a joint theme of external forces that are perceived to guide, but also – and to a higher degree – restrict and undermine the libraries’ range and extent of possible actions concerning internet regulation through filters. This includes lack of political mandate for libraries to decide on filters for themselves, being restricted by the designation in the Swedish Library Act (2013:801, 3§) of formal political mandate for library issues to the municipalities. It also includes a perceived lack of overall municipal understanding of values and requirements concerning internet library services to users.

Examples of rules commonly describe various, more or less formalised restraints imposed on libraries from their administrative authorities, the municipalities and their IT departments. Conflicts with these actors often concern the introduction of filters against the will of the library: “The library staff would have preferred that the decision-making body [the municipality] had taken the existence of the library into account.” These conflicts, however, also take the reverse form – i.e., the library wants to have filters but cannot get permission or help to install them. Other respondents state more concisely that they are not part of decisions and processes that concern the internet, and that it is common that the library is excluded from these issues. Even in several cases where it is expressed that the library is both able and willing to take part, there can be hindrances due to lack of formal grounds for or channels to exert influence: “We know what we want, but have a hard time finding the right forum, and getting on with it.” In some descriptions of problematic situations of this type, a distinction is made between the municipality in general and its IT department in particular. This often plays out to the advantage of the latter, whilst the politicians in the municipalities are described with greater scepticism: “The IT department is quite well-informed, I'm not sure about the politicians.”

Several respondents also complain about their municipality’s lack of knowledge concerning “the library’s mission”, and their “lack of understanding of the contents of the library's operations”. The problem, moreover, is not only a perceived lack of knowledge, but also that “the interest for the library’s operations as public library is low”. Responses indicate that libraries must argue for their position as unique and freestanding bodies within the municipality, and that it is problematic to clump them together with the same management and arrangements that are applied to other municipal bodies such as schools. This lack of knowledge and interest is related to a wide range of associated problems, such as a lack of concern for library-specific requirements in municipal IT procurement and libraries combatting views that “access to the internet is sometimes not even considered a required part of the libraries’ services. At least not freely”.

Lack of knowledge, however, is not merely attributed to politicians and IT departments, but is also recognised internally at the libraries: “We’re not on top of things and can’t influence that much either. The IT department almost controls us instead of being a support function.” Thus, the libraries also need more knowledge – about IT, and about the working procedures of the municipalities and their IT departments so as to find and effectively implement methods of participation in internet regulation issues.

Tools

The tools theme, finally, provides examples of libraries’ strategies to renegotiate conditions and take control of filter associated issues, even though formal power to decide on and manage this provision lies with the municipalities and not the libraries (the Library Act, 2013:801, 3 §). Tools used for such purposes include legislation and normative documents but also informal networking to educate and cooperate with local authorities and their IT departments.

Even though the Swedish Library Act does not preclude the use of filters, one of its central provisions state that the public libraries shall promote the development of a democratic society by – inter alia – contributing to the free formation of opinions (2013:801, 2§). This is often used in arguments against filters, as in this example of a library objecting to having the same filters as in schools:

The IT responsible person at our department has had discussions with the municipality’s IT department and stated that according to the Library Act the library’s computers should not have filters but offer free access to information.

Other arguments for the libraries’ rights to have special filter solutions, or none at all, provide references to library ideologies of intellectual freedom:

Upon discussions with the IT department they have presented their opinions of why booking systems, surveillance and filters are good. We have presented our users’ right to freedom of information. Our arguments had more weight.

However, no single or specific normative document was mentioned in such answers, although a general knowledge of their existence and contents may be likely sources of these arguments and positions.

The libraries’ lack of knowledge on technical and management related internet and filter issues (c.f. above) are also shown to be compensated by other, more social tools in the form of strategies to achieve good relations with, and openness and responsiveness from, local authorities: “the municipality is not knowledgeable of how we work, but we have received very good response when issues have come up.” Several respondents testify that engagement with IT departments and municipalities can give results, regardless of whether there are formal ways and platforms for this or not. Libraries that have actively invested in combatting their municipality’s and IT department’s lack of vital interest in and knowledge of library internet concerns reported that it has worked well: “We’ve worked actively to inform […] our politicians and other managers in the municipality about this [the library’s unique mission and requirements].” In all of these cases, it appears to be the libraries that are driving the contacts – and importantly, to their own benefits. But even when receptiveness for the views and requests of libraries exists, they are still commonly required to invest considerable time and resources in actively informing and spreading knowledge about themselves, their mission and the specific regulations that separate library operations from other – and at a quick glance similar – operations in municipalities.

Discussion

The first research question asked to what extent and under what conditions internet provision and use is regulated through content filters in Swedish public libraries. There are few previous quantitative studies of filter use in Swedish libraries to go by, but a comparison with the figures that can be found (Johansson and Lindh, 2020) suggests an increase in usage with about 20% in 20 years. The overall international trend also seems to be towards increased usage (ALIA, 2013; Brown and McMenemy, 2013; Kann-Christensen and Pors, 2004; Muir et al., 2016). Although difficult to compare due to differences in time, study design, and politico-administrative contexts, the results of this study indicate that Sweden’s libraries have significantly lower levels of filter use than UK and Australian libraries, and they appear to stand out through an almost non-existent use of specific filter settings for children’s computers (ALIA, 2013; Brown and McMenemy, 2013; Muir et al., 2016).

Furthermore, here the conditions of filter usage primarily refer to the manner of decision making. In this regard, the Swedish figures – whether the decisions have been for or against filters – are more clearly in line with observations from the UK generally (Muir et al., 2016) and Scotland particularly (Brown and McMenemy, 2013) which describe prominent placement of internet-related decisions and management with local authorities and their IT departments. This represents a division of labour with serious consequences for the libraries’ knowledge and decision-making mandates concerning internet provision and use. To this, we add the surprisingly high numbers of libraries that are unsure of whether their public computers are subject to content filtering or not (18%). These factual conditions described in the quantitative part of the study provide a measurable frame of reference for the qualitative analysis counterparts of libraries’ experiences and activities in dealing with aspects of these issues.

The second research question concerned what contradictions can be identified in relation to filter regulation, and the analysis revealed three main themes relating to motives, division of labour and rules and community. Starting with the motives, we can see substantial differences amongst countries concerning libraries’ attitudes to filters: frequently connoted negatively as “censorware” in the USA (ALA, 1939/1996; Buchanan, 2008; Frické et al., 2000; Knox, 2011), whilst accepted with mild reluctance in the UK (Muir et al., 2016). The Swedish libraries seem to place themselves somewhere in between these positions: they neither engage in active political activities such as public debates and policy and legislative processes like the American librarians; nor are they more or less neutral and accepting as the studies conducted in the UK describe. The local variations between Swedish libraries concerning approaches to and conceptions of filters are, however, striking. Content filters are both welcomed and requested; critiqued and resisted; and very often simply not considered at all. The main grounds for active positions nevertheless seem to be more closely related to first-hand experiences of internet use/misuse, rather than based in general principles of intellectual freedom and ethics, as also concluded for Danish librarians by Pors (2001). In comparison, perceptions of performance quality for different types and settings of filters in general (where used) appear to be of little concern.

The current Swedish situation, whereby a majority of decisions and responsibilities concerning internet regulation reside with the 290 relatively self-governing municipalities and their IT departments also raises concern regarding the second theme – division of labour. We note in relation to these tendencies an erosion of both knowledge and mandates to decide on the principal issues of filter or not, level of settings and control over daily usage such as the, according to several studies (Brown and McMenemy, 2013; Muir et al., 2016), vital rights to immediately resolve illegitimate filter blockings upon user or staff identification and request.

As the libraries are rendered powerless by a division of internet regulation labour in favour of municipalities and their IT departments, similar tendencies are at risk of being repeated between the libraries and their users, with similar consequences. Very few libraries inform their users of the existence of filters, which excludes these users from basic awareness of the conditions for their internet access and use in the library. Withholding such information may partly restrict the users’ possibilities to be mindful of potential irregularities and restrictions associated with the filters, and of the possibility to report and demand correction. The lack of such information might also be a missed opportunity to normalise a situation that otherwise causes users discomfort and even fear of being perceived as “suspicious”, making them reluctant to approach a librarian and ask them to unlock a blocked site or content (Brown and McMenemy, 2013; Muir et al., 2016).

The third theme describes restraints in the form of rules and community/actors that serve to hinder or question the libraries’ position concerning internet provision and regulation overall. The fact that such restraints to the public libraries’ operations can include local authorities’ questioning of internet provision as an integrated part of libraries’ media collections and services must be considered grave. The same can be said of tendencies to clump libraries together with IT solutions and policies for restricted municipal bodies, such as elementary schools. The national coverage of free (in the sense of gratis) internet provision in Swedish libraries today is 100% (National Library of Sweden, 2022, personal communication), but our study suggests that perhaps this figure should not be taken for granted. As public libraries still provide vital values as providers of internet to disadvantaged social groups such as newly arrived (Pilerot, 2018) and financially weak (Muir et al., 2016), safeguarding these functions may require an active defence. Nevertheless, this question is not attracting much attention in the Swedish library community.

The third research question, finally, asked how the libraries act on filter related contradictions and in what ways normative documents figure in such activities. Although the libraries in our study seem to dodge a proactive public activist agenda concerning filter use, we see reactive actions involving tools of both formal and informal character in response to perceived challenges on individual library level. These challenges can be both in the form of relatively mundane yet troublesome internet misuse experiences in the library, or as more foundational challenges to the manners and even existence of internet provision in the library from municipalities, schools, and parental organisations. Library responses to misuse are closely associated with positive attitudes towards content filters. The other, structural and political types of challenges, however, are met with both positive and negative views of filters as solutions. The overall issue of filters or not, moreover, appears secondary to a more general frustration on behalf of the libraries concerning difficulties in claiming and exercising mandate to decide on forms of internet provision and regulation on their own.

The frequent absence of formal means of collaboration and rights to be part of decision-making causes individual libraries to be more vulnerable to local variations in the willingness of municipalities and IT departments to open up information communication and cooperation. Widely surpassing references to normative documents, informal forms of collaboration emerge in our data as central tools for purposive action. These findings also suggest a positive correlation between high degrees of technical knowledge among library employees and successful outcomes of informal networking with, and “lobbying” activities towards, the municipal authorities (including their IT departments). The reliance on informal contacts and processes suggests on the one hand that there is much wanting within municipal operations, in which fair and equal treatment and formal forms of decisions are expected to be central; and on the other that a great responsibility is placed on libraries if they are to find alternative ways to establish a position of influence on decisions and daily operations.

Conclusions

The aim of this study was to explore the state of internet regulation through content filters in Swedish public libraries. We found that filter usage appears to be increasing – in Sweden, as well as in other countries. However, unless actually challenged, Swedish libraries do not seem to act on filter issues “in principle”. Even though such (in)action may appear as weakness in influential normative documents (ALA, 1939/1996; IFLA, 2012), we also see strengths and potential in a sort of context-bound flexibility that is open to practical experience and local requirements. Focusing too narrowly on the dangers and ills of filters risks increasing alternative regulatory strategies and actions in libraries, which may be “based on some of the values and norms in the profession”, but if lacking concern for related privacy issues will render the libraries’ efforts “in more dubious light” (Kann-Christensen and Pors, 2004, p. 333; c.f. also Muir et al., 2016). The filter issue is by no means an easy one, it is not a binary of “free vs censored” in a “good vs bad” way, but rather a multi-faceted and context-dependent weighting of values, ethics, legislation and local circumstances. More than anything, this further highlights the need of adequate information professional competences and mandates to decide on and oversee such regulatory decisions and management (Brown and McMenemy, 2013).

Professional normative documents were mentioned less than we may have expected as tools in situations of disagreement or conflict, but this could also be a result of the lack of explicit questions concerning this. Laws – both the Swedish Library Act and the international UN’s Rights of the Child, and Human Rights declarations were more commonly mentioned as tools in such conflicts. But above all, we notice the high importance of informal tools such as communication, networking, and lobbying towards the municipality, and the importance of IT competences for libraries to successfully influence internet regulation issues. Associated with this are disconcerting risks of individual vulnerabilities and inequalities due to specific libraries’ lack of technical and local political knowledge and connections, as well as budgetary resources to invest time and personnel in such activities. In the face of potential tendencies of an eroding knowledge base and weak and/or unpredictable formal mandates to influence, decide on and manage internet regulation, both professional and educational responses are warranted. Professional responsibility of political character primarily resides with international and national library organisations, whereas library and information science education programs should take greater responsibility for providing future library staff with a stronger knowledge base concerning information policy, internet technology and filter issues, and how to navigate and work within local political structures.

But libraries are not without responsibility for some of the inadequacies identified in this study, particularly the decision to provide users with information about filters. The surprisingly low number of respondents that inform their users about the existence – or absence – of filters is noteworthy. Why is library attention to issues of technical, information policy and infrastructural critique and transparency so blatantly rare in this context, especially given the numerous studies problematising the same? The issue appears connected to educational challenges as foregrounded above, further emphasising needs to extend attention to issues of user information and transparency as an ethical and information policy responsibility central to the libraries’ internet provision activities.

We conclude from this study that further research addressing various aspects of internet regulation in public libraries is vital. Even as internet coverage in Western or “Global North” societies is high, social divides and political instability are anything but decreasing. The ability to pursue information freely, yet in an informed balance with privacy protection considerations and user accessible information, transparency and influence options, is as vital to refugees, children, the elderly, the low-literate, the poor and adults in sensitive, violent and oppressive family situations, as to the average library user. The vulnerability of libraries to opaque and unpredictable politics on internet regulation on behalf of municipalities needs to be taken seriously. However, the current lack of studies – and the lack of coordination among them – hinders critical comparisons and context-sensitive explorations of specific social, cultural and political contexts.