Security knowledges: circulation, control, and responsible research and innovation in EU border management

ABSTRACT The knowledge emerging from research funded by the European Union (EU) through its Framework Programmes for Research and Innovation and other funding streams is significantly shaped by different forms of epistemic control exerted by the EU itself. Through the promotion of industry-research-policy cooperation in EU-funded research, and in light of the growing importance attached to ‘impact,’ this knowledge will often contribute to bureaucratic decisions taken by the European Commission, Frontex, the EU Agency for the Operational Management of Large-Scale IT Systems (eu-LISA), and other agencies tasked with border security and control. The circular dynamics surrounding knowledge production, from the calls for proposals to the results of the research, are intrinsically political and contribute to exposing the limits of the EU-promoted principle of Responsible Research and Innovation. Additionally, and due to the centrality of EU funding in the research outlook in contemporary Europe, these processes raise wider questions about the sociology of the academic fields that this article relates to: critical border studies, critical security studies, and science and technology studies. How can we interpret the interplays between the EU’s policies fostering development and integration of border security technologies, on the one hand, and the Union’s broader principles for free and open research and innovation? Through the use of autoethnographic vignettes, and mediated by an expanded Foucauldian understanding of circulation as a technology of control with performative effects, the article sheds light on the dynamics surrounding knowledge production in the field of border technologies in an EU context.


Introduction
In February 2020, the European Border and Coast Guard Agency (also referred to as Frontex) signed a collaboration agreement with the European Commission's Directorate-General for Migration and Home Affairs (DG Home) through which it will 'contribute to the effective implementation of relevant parts of the Framework Programmes for Research and Innovation (Horizon 2020 and Horizon Europe)' (Frontex, 2020a). The Horizon 2020 Work Programme for 2018-2020 had an indicative budget of EUR 118 million of EU grants for research projects in the area of 'Border and External Security'. According to the terms of reference guiding this cooperation, Frontex would, among other tasks, identify research activities addressing capability gaps in the areas of surveillance, situational awareness, biometrics, cybersecurity and information availability and exchange; translate these gaps into requirements for solutions to be researched; contribute to the development of solutions by facilitating their operational testing and validation within the framework of Frontex Joint Operations and in close cooperation with national authorities; and disseminate and exploit successful results, facilitating their market uptake and deployment (Frontex, 2020b).
In other words, and as will be demonstrated further in the article, Frontex is present at the origin of the EU border security research funding cycle (by providing input to the research programmes and respective research calls), during the development of the research (by facilitating testing and validation), and at the end of the cycle (by promoting the adoption and use of the research results). In the agency's own words, this agreement will 'contribute to the national capability development planning and the definition of the European Border and Coast Guard Agency capability roadmaps with the results and knowledge obtained from the Border Security research and innovation projects' (Frontex, 2020b). What are the implications of these circular dynamics, particularly in a field such as border security?
This article investigates how the circulation of knowledge promoted by some EU institutions and agencies has performative effects in the field of border technology, i.e. how a particular knowledge-making process leads to particular security outcomes in policy and practice. This approach enables a contestation of the material components that sustain contemporary border security in the EU.
I invoke the idiom of co-production for making sense of the dynamics where epistemic control is exerted through logics of circulation. Foucault's notion of circulation as a form of power and control provides an entry point to critically reflect on the means of control, manifesting power hierarchies and inequalities in the co-production of security-border research and policymaking. In other words, I understand circulation as co-production, i.e. as an exercise of (epistemic) power as productive to create particular socio-political orders, in this case EU border security governance. I use autoethnographic reflections to bring elements to critique how this circulation happens, as well as its consequences.
The triangle between border management, technology and security has been explored in different disciplines such as migration studies (MS), critical security studies (CSS), critical border studies (CBS), science and technology studies (STS), and international political sociology (IPS), among others. As border management strategies become increasingly reliant on new technology, it becomes fundamental to have technology at the centre of the analysis, and in particular the socio-political contexts that it emerges from, and that it impacts upon. In the context of the EU, these dynamics relate not only to the material development of border technologies, but also to wider processes of scientific and technological advancement. With the EU frameworks progressively having larger budgets, their impact on the scientific and technological landscape in Europe grows along (Felt, 2014;Stierl, 2020).
From a STS perspective, the relation between techno-scientific discourse and security has received continuing attention, especially through a focus on how security matters are intimately bound up with how we might 'know' about social and technological complexities and risks inherent in large-scale technical systems (Vogel, 2017, p. 977;Aradau and Perret, 2022). This is particularly salient in the field of border management, where border security 'problems' and respective 'solutions' are co-produced by complex processes involving a multitude of actors (Jasanoff, 2004;Trauttmansdorff, 2022). In this article, I understand the process of knowledge circulation as a form of co-production. It is the knowledge about border security that orders the governance of border security.
Theoretically and conceptually, this article aims at 'building upon generative intersections' (Suchman, 2020, p. 1) between STS and CSS (see also Evans et al., 2020), advancing avenues opened by the work already carried out in the field of border control and security technologies, including Bourne, Johnson and Lisle on laboratizing the border (2015), Baird on border security fairs (2017) and Bigo on practices and everyday security (2014). All these perspectives have brought important contributions to a contestation of the border Vaughan-Williams, 2009, 2012) and the security politics associated with its management.
The article has two objectives. The first one is to show the relations between the concepts of co-production, circulation, and RRI. This is done with a conceptual discussion and with an empirical analysis of data generated from public sources. The second objective is to critique and interrogate those same relations within a sociology of knowledge production in Europe (Law and Mol, 2002), centring issues of positionality of researchers, the role of EU research funding in the European academic job market, and the wider possibilities for critique emanating from this scenario. This will be done departing from autoethnographic vignettes and a self-reflection surrounding my own engagement with these processes.
Methodologically, then, the article combines open document sources with an autoethnographic approach to illustrate, on the one hand, how dynamics of circulation in border security technological knowledge and expertise co-produce the field of EU border security, and, on the other hand, the consequences this has in terms of challenging the principles of open science and RRI.
The issues raised in this article continue the discussion opened in the pages of Science as Culture with the forum on embedding social sciences and humanities (SSH) and their role in performing wider research agendas. The article, and this special issue as whole, bring elements to contest 'how do and can SSH intervene in technoscientific research policy and priorities' (Felt, 2014;Levidow, 2014, p. 382;Levidow and Neubauer, 2014). This effort is part of a growing self-reflection from STS, CSS and CBS scholars about their positionalities and how critique remains possible in the context of engagement with communities of practice (Austin et al. 2019;Leese et al., 2019;Bellanova et al., 2020;De Goede, 2020a).

Analytical perspectives
The circulation of knowledge as a technology of border security governance Over the last two decades, the notion of circulation as a technology of control has been used by scholars in CSS and CBS to explain the transformation of the articulations between power, security and society (Salter, 2006;Dillon and Lobo-Guerrero, 2008;Aradau and Blanke, 2010;Wichum, 2013). This approach to the concept of circulation was developed by Michel Foucault in his 1977-1978 lectures on Security, Territory, Population (Foucault, 2007) and it was used to propose a new understanding of securityfor Foucault, security referred to the biopolitical practices of 'organizing circulation, eliminating its dangers, making a division between good and bad circulation, and maximizing the good circulation by eliminating the bad' (Foucault, 2007, p. 18). With an understanding of population as the referent object of security instead of the political entity that governed itthe notion of circulation is particularly apt in explaining mobilities and the politics associated with its governance, which in turn has made it relevant for the study of borders (Salter, 2006(Salter, , 2012Johnson, 2011;Walters, 2011;Leese, 2016).
In this article, I draw on the notion of circulation as a technology of power, in the sense that controlling the circulation is an exercise of power. Yet, contrary to Foucauldian-inspired branches within critical security studies that have looked at the circulation of people, I am interested in the circulation of ideas and knowledge, and how these are generative of new social and political orders. In this sense, circulation can be understood as a form of co-production, since it engages with the processes by which 'knowledge-making is incorporated into practices of state-making' (Jasanoff, 2004, p. 3). In the case of the border technology, I highlight how the processes of scientific and technological knowledge creation exhibit a logic of circulation that has material consequences.
Indeed, circulation can be interpreted as a wider cultural phenomenon that impacts the non-material world and constitutes a form of co-production of social and political orders. For Benjamin Lee and Edward LiPuma, circulation is 'a cultural process with its own forms of abstraction, evaluation, and constraint, which are created by the interactions between specific types of circulating forms and the interpretive communities built around them' (Lee and LiPuma, 2002, p. 192). In this sense, circulation of knowledge has performative effects and is crucial for producing both material artefacts (Mackenzie, 2005;Callon, 2006) and the social and political orders that govern them.
This fluid understanding of knowledge as an ordering movement rests on the idea that the processes through which contemporary societies form their epistemic and normative understandings of the world are co-produced (Jasanoff, 2004). But, as highlighted by Jasanoff, co-production should not be advanced as a fully-fledged theory, claiming lawlike consistency and predictive power. It is far more an idioma way of interpreting and accounting for complex phenomena so as to avoid the strategic deletions and omissions of most other approaches in the social sciences. (Jasanoff, 2004, p. 3) My argument is that the dynamics of circulation surrounding knowledge and material production in the field of border security in the EU are both key to the ordering practices governing border security and, relatedly, a form of epistemic control with clear material and political repercussions. As Herberg and Vilsmaier have argued, drawing on both classic and more recent studies, there is a 'critical conceptual and historical connection between scientific and political power,' and therefore 'epistemic and social control is co-constituted' (Herberg and Vilsmaier, 2020;p. 309. See also Callon and. This special issue interrogates how border control technologies turn 'borders' into sites of contestation over different and sometimes conflicting policy agendas, policies and measures around migration. Its critical reflection on the material artefacts that constitute borders enables a political contestation of the processes that take place before that material constitution. I follow Claudia Aradau's reflection on the function of critique in the field of technology: 'while technologies of (in)security often appear different, novel, unprecedented, critical work needs to reformulate analytical tools that can grasp the reconfiguration and recomposition of discourses, technologies and practices' (Aradau, in Salter, 2019, p. 24). From this perspective, it becomes relevant to analyse the processes by which the material artefacts come to being, and this article speaks directly to that aim.
The circulation of technological border security knowledge in the EU How does border security knowledge emerge in the EU, and who defines its gaps? In recent years, social science research has focused extensively on the growing deployment of technology to the EU borders, as well as on the construction of digital and surveillance infrastructures. But the processes by which those material realities come to being are less known. This article builds on the lines of inquiry proposed by Bourne et al. (2015) that interrogate 'how borders are produced by scientists, engineers and security experts in advance of the deployment of technical devices they develop ' (Bourne et al., 2015, p. 307). As I will demonstrate, important material elements of contemporary EU border security technologies have its origin in knowledge emerging from research funded by the EU itself, following a process that was formally and substantively initiated by the EU itself through detailed calls for proposals.
The circular movements involving border security research and border security materialization means that the EUthrough its many different bureaucraciesenacts a crucial role in governing knowledge production in the field of European border security, which in turn contributes to the materialization of border security itself. In other words, the different interventions of EU agencies in the research process cycle, from the research call to the promotion of the 'solutions' resulting from that research, create a logic of circulation that perform the agency's political agenda, generating material expressions of that same agenda in the form of technological artefacts.

Responsible research and innovation
While border security research is one of several thematic areas in the EU research programmes, there are some ideas that are horizontal across the different areas. One of the most important principles is Responsible Research and Innovation (RRI). In the European Commission's definition, RRI 'is an approach that anticipates and assesses potential implications and societal expectations with regard to research and innovation, with the aim to foster the design of inclusive and sustainable research and innovation' (European Commission, 2020a).
RRI 'emerges from the recognition of the power of science, and this recognition has forced reconsiderations of the responsibilities that should follow such power' (Stilgoe and Guston, 2017, p. 853; see also Burget et al., 2017; for RRI in the domains of security, see Burgess et al., 2018 andAhmad, 2020). In order to fulfil this task, the Commission requires the principle to be implemented as a package that includes multi-actor and public engagement in research and innovation, enabling easier access to scientific results, the take up of gender and ethics in the research and innovation content and process, and formal and informal science education. (European Commission, 2020a) The recognition of the power of science and knowledge is important for our argument that the circulation of knowledge requires critical engagement due to its inherently political character. RRI implies that the recognition of such power has forced reconsiderations of the responsibilities that it entails. Indeed, Owen, Macnaghten and Stilgoe argue that there are three main discourses associated with the idea of RRI: an emphasis on the democratic governance of the purposes of research; the idea of responsiveness (emphasizing established approaches of anticipation in research and innovation); and the framing of responsibility itself in the context of research and innovation as collective activities with uncertain and unpredictable consequences (Owen et al., 2012). In the words of the Commission, 'RRI is a key action of the 'Science with and for Society' objective' and its actions will be promoted via the 'Science with and for Society' objective' (European Commission, 2020a).
Yet, RRI is also an intervener in the co-production dynamics at stake in this article. As highlighted by Nina Klimburg-Witjes and Frederik C. Huettenrauch, 'RRI itself embodies a political agenda, conceals alternative experiences by those on whom security is enacted upon and that its key propositions of openness and transparency are hardly met in practice due to confidentiality agreements' (Klimburg-Witjes and Huettenrauch, 2021, p. 1). In which way are these principles challenged by security research, particularly the one that deals with border security? And how can we investigate that?
Methodologic reflections: from autoethnography to 'staying with the trouble' In order to carry out research on this topic, I draw on empirical elements generated from both written sources and reflections upon my own engagements with the processes under study in a retrospective autoethnographic exercise (Ellis et al., 2011) using autoethnographic vignettes. I follow Humphreys in understanding 'autoethnographic vignettes as an alternative approach to representation and reflexivity in qualitative research' (Humphreys, 2005, p. 840), which in this article illustrate first-hand encounters with/of the dynamics that I research in the same article. In the words of Denzin, vignettes 'ask readers to relive the experience through the writer's or performer's eyes' (Denzin, 2000, p. 905). As explored by Marieke De Goede, vignettes are 'understood as short narrative stories or examples' that work as a 'window onto the practical and ethical dilemmas of studying (…) security practices' (De Goede, 2020b, p. 262), creating a space for self-reflection and critical introspection. De Goede specifies the conditions when vignettes are a useful tool in security research, suggesting that, first, 'vignettes exemplify the mundane, everyday, lived reality of research practice' (De Goede, 2020b, p. 262), and, second, 'vignettesor auto-ethnographic interludesare particularly useful to make space for accounts of how secrecy was negotiated and navigated in fieldwork' (De Goede, 2020b, p. 263).
This aspect is relevant for this article, since the processes analysed here are often not only surrounded by secrecy related to the security component of the 'solutions' (methodologies and technologies) developed, but also by secrecy around the elaboration of the calls, the award criteria, and also the fact that many of the deliverables produced in the projects are not public. We therefore follow de Goede in interpreting vignettes as a particularly useful tool in spaces of secrecy, and this is part of the added value they bring to the article.
Whereas some argue that narrative vignettes should be based on fieldnotes taken as the events happened (for example Humphrey 2005), my vignettes are retrospective reflections of lived experiences, capturing my contemporary thoughts on past events. Yet, these thoughts are also a product of oral discussions and email exchanges with colleagues held at the time those events happened. My experiences were generated out of different ethnographic sites. First, these include years of participation in different projects funded by the EU within the security research programme, 1 as well as many applications to different Horizon 2020 and Horizon Europe calls. This gave me first-hand insights into the logics put into the formation of the consortia that apply to the funding, the drafting of the proposals in relation to the calls, the actual conduct of the research, the different strategies for dissemination and communication with stakeholders and end-users, as well as post-project legacy activities. Secondly, research for this article builds on attendance of industry exhibitions in the US and in Europe, as well as observation of meetings between technology industry and border guards. Thirdly, it benefits from my talks and exchanges with civil society organizations that monitor developments in EU border security technologies, including Statewatch, Drone Wars UK, PAX for Peace, and the European Forum on Armed Drones. Finally, my former position as short-term political analyst within an EU institution, 2 as well as my engagements 3 with different EU institutions across time, enable an understanding of the logics permeating research and policy as seen from an EU perspective.
This multitude of experiences form my position in the research field and situate the knowledge that I have and the questions that I choose to investigate. As developed by Donna Haraway, the notion of situated knowledges has different dimensions: epistemological, ontological, ethical, and political (Haraway, 1988). The epistemological and ethical, in particular, are key for the definition of the questions raised in this article: an awareness of this positionality is crucial for the process of acknowledging how my embeddedness in some of the processes described here impacts the questions that eventually I raise. In the case of this article, I opt by 'taking the trouble,' i.e. acknowledging the 'need to stay close to the troubles that new research objects pose to the study of security' (Bellanova et al., 2020; see also De Goede, 2020a).
As EU-funded research becomes a growing part of an increasingly job-insecure European academia, the politics associated with a funding entitythe EU that has political and security agendas in all sorts of areas beyond science needs to be discussed. As recently asked by Maurice Stierl in the context of an analysis of the consequences of EU-funded migration research, 'what is the responsibility of researchers for the knowledge they produce and disseminate?' (Stierl, 2020, p. 1; see also Gläser and Velarde 2018), particularly considering that today research is often required to inform policy decisions and have 'an impact.' This leads to wider questions, since much of the research promoted by the EU through its framework programmes has several other objectives beyond basic science or the production of particular solutions to a specific problem (see for example Martins and Kuesters, 2019, on how the Framework programme 7 and Horizon 2020s security research programme aimed at creating a European culture for security R&D).
The theoretical and methodological reflections presented so far will later be called back to shed light on the empirical material of the article, which is made of two main components. The first uses open-sourced texts of the EU calls, Frontex documents, consortium websites and newsletters related to Frontex' engagement with research carried out pursuant the Border and External Security funding stream of Horizon 2020s Work Programme 2018-2020 (SU-BES01-2018-2019-2020), point 14 -Secure societies: Protecting freedom and security of Europe and its citizens. The second component is made of different autoethnographic vignettes reflecting on my years of engagement with EU-funded research.

EU-funded border security research
Since the 2004 Hague Programme, and with an impetus from the Smart Borders Initiative introduced in 2011, higher levels of technologization have been portrayed in EU policy documents and strategies as a panacea for the Union's border management problems (Jeandesboz, 2016;Leese, 2016;Rijpma, 2017). However, while the literature has certainly looked into this process of equating increasing technologization with increasing levels of border security (Dijstelbloem and Meijer, 2011;Lehtonen and Aalto, 2017;Kloppenburg and van der Ploeg, 2020;, it has not paid enough attention to socio-political dynamics involving the production of the knowledge upon which decisions are made (relevant exceptions include Bigo, 2014;Baird, 2017;Binder, 2020;Klimburg-Witjes and Huettenrauch, 2021;Aradau and Perret, 2022). As this section will show, much of this knowledge emerges from research funded by the EU itself through its framework programmes and other funding streams, where the content of the calls, as well as the monitorization of the consortia carrying out the awarded projects, significantly shape and frame the content of the knowledge produced. This knowledge will in turn contribute to bureaucratic decisions taken by the European Commission, Frontex, EMSA and other agencies tasked with border security and control.
Border security has been a very important part of the security research programme (SRP), one of the research lines introduced by the EU Framework Programme for Research and Innovation 7 (FP7) that run from 2007 to 2013. In FP7, the border security priority was presented as aiming at focusing on technologies, equipment, tools and methods for protecting Europe's border controls such as land and coastal borders (see Bigo et al., 2014). Its successor, Horizon 2020, expanded the EU's commitment to funding security research, and border security remained a top priority. As explicitly stated in the programme's webpage, one of the primary aims of the Secure Societies Challenge (the official name of the SRP in H2020) in the field of border security is to improve border security, ranging from improved maritime border protection to supply chain security and to support the Union's external security policies including through conflict prevention and peace building ( Table 1).
To these priorities corresponded many different calls. Out of these calls, dozens of different projects were awarded funding to conduct research that aimed at delivering the results envisaged by the calls. A crucial component of the SRP is its focus on technology development: the programme has the 'expectation of both generating new knowledge and promoting the application of new technologies while reinforcing the competitiveness of the European security industry' (Martins and Ahmad, 2020, p. 65). The projects materialize the triple helix model of innovation (Leydesdorff and Etzkowitz, 1998), referring to the interactions between academia, government and industry, designed to foster economic and social developmentand in more recent times, increased civil society participation became an objective as well. To put it differently, the consortia created to answer the calls typically involve research institutions (universities, independent labs and private research institutes), governments (normally law enforcement agencies, border security forces and security ministries, in the case of the SRP) and the industry (spanning from start-ups and SMEs to Europe's larger defence industries).
It is important to mention that not all EU-funded research goes through the same procedure. For example, the grants attributed by the European Research Council, the Jean Monnet modules, or the Marie Skłodowska-Curie Actions, for example, are awarded strictly on the basis of academic excellence. The proposals submitted by the researchers in these schemes do not respond to any specific call where the content and objectives are very much pre-determined. The processes described in this article, therefore, are not observed in those funding schemes, and report only to the research funded under the SRP, that has manifestations in FP7, in H2020, and in the current Horizon Europe programme as well.

Circulation in EU-funded border technology research
How can we account for the materials that shape and produce border security technology knowledge in the EU, and how do they form the conditions for the above-mentioned dynamics of circulation? In the next paragraphs, I will illustrate the latter with processes surrounding the most recent EU framework funding for research on border technologies. This was part of Horizon 2020s Work Programme 2018-2020, point 14, called Secure societies: Protecting freedom and security of Europe and its citizensthe official designation of the SRP. Under this programme, there was a funding stream called Border and External Security with three topics: (1) Human factors, and social, societal, and organizational aspects of border and external security; (2) Technologies to enhance border and external security; and (3) Demonstration of applied solutions to enhance border and external security. Each of these topics had, then, different sub-topics, to which corresponded specific calls (see Table 2). The process that leads to the formulation of the specific calls is widely opaque, but Frontex has confirmed that it provides inputs to these calls. In the context of the above-mentioned Terms of Reference framing the 2020 cooperation agreement between Frontex and the European Commission, Frontex will 'identify research activities addressing capability gaps in the areas of surveillance, situational awareness, biometrics, cybersecurity and information availability and exchange' and 'translate these gaps into requirements for solutions to be researched' (Frontex, 2020a). In other words, the calls for border security research in Horizon 2020 (and its successor for 2021-2027, Horizon Europe) will reflect specific operational and capability gaps of Frontex, the same applying to other agencies and systems such as EUROSUR. Additionally, the calls include mentions to different policy agendas and strategic priorities, such as the EU Serious and Organised Crime Threat Assessment 2017 (European Commission, 2020b, p. 58) or the Smart Borders package (European Commission, 2020b, p. 60). This indicates that the calls for proposals are, often, at the service of different EU policies. In the words of Clemens Binder, since the framework programmes pursue a 'policydriven' approach, these calls often seek to address specific political goals (Binder, 2020, p. 148).
When the projects eventually start, and the consortia are formed, Frontex remains engaged in some of their activities. These include tests, trials and demonstrations of technologies (Frontex, 2020c), but also participation in workshops and interim reviews of some projects, which are key milestones in the projects' workplans. This is what happened, for example, in January 2021, when Frontex joined the Research Executive Agency (REA), the European Commission's Directorate-General for Justice and Home Affairs (DG Home) and different national authorities to review the ARESIBO project (Frontex 2021). This project aims at 'improving the efficiency of the border surveillance systems by providing the operational teams and the tactical command and control level with accurate and comprehensive information,' using augmented reality and different unmanned vehicles, and received EUR 7 M for the period 2019-2022 (Cordis, 2019). When the projects are concluded, Frontex participates in the evaluation of the results and promotes the use of the 'solutions'. According to an August 2020 news note, the agency has identified seven flagship projects emanating from the calls enunciated above in Table 2: Andromeda, ARESIBO, BorderSens, COMPASS2020, D4FLY, MIRROR and PERCEPTIONS (Frontex, 2020b). These projects provide different technological 'solutions' to gaps in border management previously identified by EU agencies themselves. For this reason, the closer their proposed 'solutions' are to those gaps, and the more integrated they are with EU and national policies and regulations, the more valued they are by the EU agencies. The Andromeda project, for example, claims to be 'perfectly aligned with the overall European policy that facilitates the interagency interoperability and cooperation and allows each Member State to decide how, when and whether additional data sources are of relevance to its operations' (Andromeda, 2021). The BorderSens project, which combines 'robust sensor technologies with the inherent advantages of electrochemical strategies, nanomolecularly imprinted polymers, and multivariate and pattern data analysis,' has a 'straightforward exploitation route ' (BorderSens, 2021) to accelerate the adoption of their proposed 'solutions' by border authorities.
The technological 'solutions' these projects propose are very diverse, and some projects develop many different systems based on different technologies. The D4FLY project alone, for example, proposes the following proud menu of techno-solutionism: Biometric technologies for the verification of traveler's identities on-the-move through 3D face, iris, and somatotype recognition; Smartphone applications for enhanced traveler verification; Thermal and multispectral imaging for counter spoofing; Advanced morphed face detection algorithms through Convolutional Neural Networks; Deep Neural Networks and advanced methods for impostor and document fraud detection; Computer vision algorithms for automated passport forgery detection. (D4FLY, 2021)

Contesting border security knowledge
What are the consequences of these circular dynamics surrounding knowledge production in the field of border security? The main outcome of these processes is that many of the technologies that constitute the material enactment of the EU border ultimately derive from pre-established political agendas, and not from basic, independent research. In other words, Frontex and its policy agenda do not merely constitute an inspirational or analytical actor in conceiving and implementing these projects. Frontex actually performs its agendas through these programmes and its engagement with the projects. Importantly, these political agendas are, as demonstrated above, mediated through technologies. But where do the ideas for creating these technologies come from? Vignette 1 illustrates how the processes of circulation are triggered. Vignette 1triggering circulation In a conversation at the Frontex headquarters in Warsaw in February 2018, a senior official told me that 'the practitioners never ask for the technologythey are too busy for that.' He further explained that there's never a direct demand from the practitioners; it is, he argued, 'only the industry and the academy that develop ideas and products and they do it because it is their business.' The industry has two product pullers: their own strategy departments, which survey the markets and make prospective analysis of tendencies and needs; and tenders and calls from funders and future potential end-users. As I listened to this explanation, it became clear to me that these two product pullers do not exist in isolation; rather, market trends and (imagined) future needs also constitute the knowledge that shapes the content of the EU calls and tenders.
In this case, we see the logic of co-production on display. It is the knowledge that orders governance, and, in the case of this article, knowledge is about framing something as a security issue amenable to a technological solution. This is problematic on many levels, considering the growing criticism surrounding the agency's actions when it comes to border management. Over the last two years, a wide array of political actors has called for the resignation of the agency's director Fabrice Leggeri or have demanded higher levels of protection of fundamental rights to migrants. On 1 May 2022, eventually, Mr. Leggeri resigned after investigations that the agency had been involved in illegal pushbacks of asylum seekers attempting to reach EU territory. The actors calling for his resignation included different control agencies (such as the EU Ombudsman, for not suspending operations on the Aegean, or the European Anti-Fraud Office, for a case of possible fraud related to the agency's IT contracts), members of the European Parliament (through approving a February 2021 report calling for respect for human rights at the EU's external borders, entitled 'Implementation of Article 43 of the Asylum Procedures Directive'), the European Commission (in the case of documented direct and indirect Frontex involvement in pushbacks from Greece in December 2020), and countless NGOs. Different cases are pending in EU courts involving Frontex, in cases following charges presented both by civil society against Frontex and Frontex against civil society organizations. That research carried out in universities and research institutes is part of the activities of this agency is something that opens some questions and merits self-reflection from researchers. That some of that research was funded in a scheme that advocates the principles of responsible research and innovation requires further scrutiny.

Contesting and deconstructing 'solutions'
The point relating to the performative character of the circulation dynamics is part of the broader issue of epistemic control. The dynamics identified above put research at the service of pre-established research goals. By defining gaps a priori, Frontex and other participants in the drafting of the calls performatively enact a set of assumptions about what constitutes a security problem. This performativity produces material effects, first in the form of calls, and ultimately in the form of the 'solutions' to the predefined 'problems.' As shown by Martins and Jumbert, in this process security 'problems' and 'solutions' are coproduced (Martins and Jumbert, 2022). Part of the process of reaching an understanding about what constitutes a knowledge need (and what could be a solution for that) results from practices of socialization that take place between technology developers, policy-makers, and potential end-users of that technology. Vignette 2 illustrates how some of these dynamics take place in the field of EU border security.
Vignette 2 -Spaces and practices of socialization Over the last seven years I have attended multiple events in which representatives from the tech industries meet with stakeholders from law enforcement agencies, border guards and other public authorities. The settings in which these events are organized can have multiple formats: technology exhibitions in which a number of lectures / talks / roundtable discussions take place; specific close-door meetings where selected technology providers show-case their products to key decision-makers; symposia where different actors make presentations to a specialized audience; exercises and technology demonstrations in simulated and/or real scenarios. 4 Irrespective of the format these events take, the opportunities for socialization and informal conversations are always present and constitute a crucial part of the agenda. There are ample coffee and lunch breaks, precisely to facilitate these exchanges. In one of the meetings I attended, there was an agenda item called 'The Mix,' where some drinks and finger food were served and the explicit objective was to have people from different areas of practice to mix, mingle, and informally chat. There is nothing spontaneous in these 'spontaneous' informal talks that take place during these events. Herberg (2020) has persuasively argued that, in collaborative research, 'phase zero' (i.e. the very beginning of the collaboration) is not co-designed but scripted: 'The early moments of collaborative research are entwined with a tacit, tactical, and relational form of control' (Herberg, 2020, p. 395). This scripted character of some types of research is, in the case under analysis in this article, further enhanced when the calls are embedded in pre-defined objectives, goals, and politics. What eventually results in a 'solution' at the end of an EU-funded project is much less the result of free, open research than a guided path across a scriptand for this reason, I would argue, the word 'solution' needs to be contested. By contributing to defining what is to be known, the calls fundamentally enact epistemic control over the border security research that is to be funded by the EU.
In the same vein, there are cases in which a big part of the technological infrastructure for border security is already in place, and the funding available is meant only to make those different systems interoperable and compatible. Vignette 3 illustrates one of those cases. excludes, In 2021 I was approached to join a consortium that was putting together a proposal answering a Horizon Europe's Border Management call. 5 This invitation was aimed at having my institution as a project partner in the proposal. This was a technological project (i.e. with the technology at the center) on maritime surveillance, created for developing new operational concepts to improve coordination and interoperability of multi-level platforms (with a strong focus on unmanned vehicles), with the final aim of improving surveillance at EU borders. As usual, the partners were industry, a member state navy, and applied research institutes. Importantly, the different technological platforms (i.e. the vehicles) and payloads (e.g. SIGINT sensors and optical scanners for drones) already existed individually, so the project would be about developing ways of coordinating them. Due to its high budget, participating in the project would enable my institution to achieve relevant financial gains and perhaps would lead to new hirings of academic staff.
Following internal discussions, we decided not to join the proposal, since the potential for impact of our role as researchers with experience in research ethics and societal impacts of security technologies would be marginal. The latitude for us to exert real influence on what seemed to be a project already consolidated would be minimal.
Vignette 3 illustrates two important dynamics. Firstly, the idea that the budgets involved in security research are often relevant. In an academic profession marked with job insecurity, the availability of relevant funding is a material factor that cannot be ignored. Secondly, this case also shows that the possibility for critique within the consortia can be absent (or nearly so). While there are many possibilities for meaningful engagement in EU security research from researchers that work on research ethics and societal impacts of security technologies, there are some contexts where their potential for impact is largely absent. This conclusion resonates with the arguments already put forward by Martin-Mazé and Perret (2021) and by Leese et al. (2019), when they argued that applied ethics in EU-funded security research 'face considerable challenges that result from its location in the middle of numerous cross-pressures, such as political ambitions, economic interests, technological rationales and the demands of security professionals' (Leese et al., 2019, p. 59). These cross-pressures create a dynamic that include some actors, exclude others, and leaves others in-between.
It is also important to underline that, from the perspective of some EU officials, the circulation of knowledge is not only desired, but indeed needed. Enhancing circulation is explicitly portrayed by the EU as a necessity and a logical measure in order to exponentiate the material and non-material dividends of the EU investment in knowledge production. Vignette 4 illustrates the logic behind this thought.
Vignette 4 -Circulation as a desired outcome Due to my participation in different consortia funded by EU programmes (FP 7, Horizon 2020 and Horizon Europe), I am aware that the intellectual property rights of the research carried out in the project rest in the consortium. Article 16.2 ('Ownership of results') of the Horizon Europe Model Grant Agreement states that 'the granting authority does not obtain ownership of the results produced under the action.' While researchers see this provision as beneficial for their career development, some EU officials are critical of it, seeing it as limiting the possibilities for exploiting the research results, given that normally the consortia dissolve after the completion of the project and the results get dispersed or lost. I was told by an EU official with a border security portfolio that 'in the EU what they do is to put together a puzzle that is dismounted after the demonstration.' Many (if not most) of the technological solutions developed in the consortia for border security neither get market traction nor are adopted by end-users. For this reason, some effort has been put by stakeholders in EU agencies in working on the so-called projects legacy, i.e. involving the potential end-users of the 'solutions' developed by the consortia at all stages of the project life, particularly the last ones.

Contesting responsible research and innovation
On the basis of the empirical material presented, I argue that these processes challenge the EU-promoted principles of RRI. The notion of RRI implies that 'societal actors (researchers, citizens, policy makers, business, third sector organizations, etc.) work together during the whole research and innovation process in order to better align both the process and its outcomes with the values, needs and expectations of society' (European Commission, 2020a). In 2014, the Commission had already published the report Responsible Research and Innovation: Europe's ability to respond to societal challenges, in which it presented the six pillars of RRI: 'choose together,' 'unlock the full potential,' 'creative learning, fresh ideas,' 'do the right "think" and do it right,' 'share results to advance' and 'design science with and for society' (European Commission 2014).
My argument is not that Frontex or other EU operational agencies in their everyday operational activities should abide to RRI, which is a principle that applies to R&D activities. But the practice exhibited by the Commission and other EU agencies, in particular through its funding of border security technologies, raises challenges to the framework. One key aspect has to do with public engagement and access to research results. Both the Parliament and other EU agencies have difficulties in getting from Frontex the necessary information about some of its activities. Additionally, some of the most ethically problematic border research funded by the EU in the field of border security has led the Parliament to demand further information from the Commission, in the latter's capacity as funding institution. Yet some of these requests for information were not sufficiently addressed by the Commission (Hayes et al. 2014;Martins and Kuesters, 2019).
The agenda-setting powers of the calls also impact principles of basic research and challenge the fluid character of knowledge production. With stakes involved in all stages of the research cycle, Frontex exerts epistemic control over an issue that has severe human rights implications of vulnerable people. That a law enforcement agency with growing military-type assets exerts that epistemic control should not be acceptable.
Contesting the field: situated/circular knowledges This article's methodological and theoretical options provide an opportunity for critical self-reflection from the STS, CSS and CBS communities. The sociology of the field is severely marked by the agenda-setting powers of research funders, and these powers grow in parallel with the amounts of research funds provided. As shown here, research agendas are commonly performed a priori by the funders at the service of policy agendas, and only a posteriori met by the proposals that researchers design in order to match the original calls. This is especially observable in the case of EU security research funding, whose volume has been growing for 15 years and where border securityin particular technology-centred research has been a central part. While an investment in research is certainly welcomed, it nevertheless raises questions about its stated priorities, the content of the specific calls, and the expected output of that same research. In the case of the topics covered in this article, the funds available for border management research have grown steadily over the last two decades and, in parallel with the impact on security practices explored here, they also had a clear impact on the security and migration studies landscape in Europe (Stierl, 2020). Owen et al. (2012) argue that one of the main discourses associated with the idea of RRI is an emphasis on the democratic governance of the purposes of research. The close involvement of Frontex, a law enforcement agency, in the different stages of the research cycle, has consequences for the observance of RRI principles.
In an article addressing issues of policy-relevance of security research, Harmonie Toros expresses concern that 'the famous "so what" question asked about any research has gone from meaning 'how does this contribute to knowledge?' to 'how does this contribute to knowledge and how can it have relevance beyond academia, including in the policy world?'' (Toros 2016, p. 126). The fact that research came to be equated by some funders with policy impact presents challenges to intellectual freedom and to the conduct of basic research.

Conclusion
This article has showed how, in the context of EU-funded border security research, Frontex and other agencies are often present at all the stages of the research funding cycle: at its origins (through input to the research programmes and respective research calls), during the development of the research (through facilitation of testing and validation and through participation in the consortia), and at the end of the cycle (through the promotion of the adoption and use of the research results). The knowledge emerging from EU-funded research is, therefore, significantly shaped and framed by the content of the calls issued by the EU itself. This knowledge eventually contributes to bureaucratic decisions taken by the European Commission, Frontex, the European maritime Safety Agency, and other agencies tasked with border security and control. These circular dynamics are a form of co-production and expose the limits of the RRI framework that the European Commission demands from all R&D it funds. Additionally, these processes raise wider questions about the sociology of the academic fields that this article relates with.
What are the implications of these circular dynamics, particularly in a field such as border security? The concepts of circulation and co-production shed light on different aspects of knowledge production in the field of border technologies in an EU context. The dimension of epistemic control, and its impact on the observance of RRI, enabled a critical look into the functioning of EU-funded research in the field of border security, a field where policies, strategies and priorities have life-changing/threatening impact on highly vulnerable people.
The situated knowledge emerging from this research raised questions about the place and the role of the researcher in these processes. The article indeed raises questions, but, it is acknowledged, does not offer many solutions. While there is nothing intrinsically problematic with EU-funded research aimed at providing research-based solutions to world problems, quite the opposite, it is important that researchers have their own high, critical, self-reflexive standard for accepting partaking in collaborative research. If and when some aspect of the project raises questionsbe it the angle of the research, the methodology employed, or the profile of an institutional partner, for exampleit should be confronted; we should 'take the trouble' (Bellanova et al., 2020), confront it, and try to solve it. RRI is also about our own standards, about what kind of researcher we want to be, and what kind of contribution we want to make to the community. As argued by Marc Steen, responsible innovation requires time and space for selfreflection, and it is therefore necessary to 'make time for uneasy questions, vulnerable experiences, awkward moments and uncertainty' (Steen, 2021, p. 254).
While the prospect of a successful grant application that can bring job security is something that needs to be acknowledged, our duties as responsible researchers and innovators, as members of the research community, cannot be neglected either.
By bringing this issue to the frontline of its inquiry, the article engages with the critique, contestation and reflexivity that sits at the core of the fields of STS and CBS. Its focus on the sociology of knowledge production relates to what Vogel et al. call reflexing knowing (2017, pp. 985-989), referring to the multi-layered tensions for scholars investigating security matters, and how STS's own knowledge-making practices are shaped by its engagements (Vogel et al. 2017, p. 974). I have argued that Frontex' participation in the different stages of the research cycle in EU-funded border security research constitutes a form of epistemic control and is representative of larger processes at stake. To put it differently, Frontex and its policy agenda do not merely constitute an inspirational or analytical actor in conceiving and implementing these projects; Frontex in fact performs its agendas through these programmes and through its engagement with the projects. The political character of these processes requires further attention and contributes to the contestation of how borders are created and enacted, and what constitutes border security 'problems' and 'solutions.' In this sense, this article is, in the words of Johnson et al. (2011), an intervention on rethinking the border.

Notes
His work addresses critical issues at the intersections of technology development, security practices, and societal change.