Search

The adoption of Artificial Intelligence (AI) by the public sector has the potential to improve service delivery. However, the risks related to AI are significant and citizen concerns have halted several AI initiatives. In this paper we report findings from an empirical study on citizens´ attitudes towards AI use in public services in Norway. We found a generally positive attitude and identified three factors contributing to this: a) the high level of trust in government; b) the reassurance provided by having humans in the loop; c) the perceived transparency into processes, data used for AI models and models´ inner workings. We interpret these findings through the lens of social contract theory and show how the introduction of AI in public services is subject to the social contract power dynamics. Our study contributes to research by foregrounding the government-citizen relationship and has implications for public sector AI practice. Paid open access

The adoption of Artificial Intelligence (AI) by the public sector has the potential to improve service delivery. However, the risks related to AI are significant and citizen concerns have halted several AI initiatives. In this paper we report findings from an empirical study on citizens´ attitudes towards AI use in public services in Norway. We found a generally positive attitude and identified three factors contributing to this: a) the high level of trust in government; b) the reassurance provided by having humans in the loop; c) the perceived transparency into processes, data used for AI models and models´ inner workings. We interpret these findings through the lens of social contract theory and show how the introduction of AI in public services is subject to the social contract power dynamics. Our study contributes to research by foregrounding the government-citizen relationship and has implications for public sector AI practice. What do citizens think of AI adoption in public services? Exploratory research on citizen attitudes through a social contract lens

The energy industry in Norway has a long tradition in using information technology to enable integrated operations, namely, remote collaboration between personnel at offshore installations and experts at onshore office environments. Currently, the industry is undergoing a digital transformation in which remote operations of unmanned offshore assets are the emerging standard. To ensure trustworthy and reliable operations, offshore remote sensing capabilities must be established through not only technical means but also a broader transformation involving new competence, work processes, and governance principles. In this chapter, we reconstruct this transformation and ask: What are the emerging capabilities that develop around the remote operation digital infrastructure? We unpack how the new digital infrastructure is a continuation of the practices and systems that have been established over time. We use historical reconstruction with vignettes from the development of a new generation of remotely operated offshore installations in oil and gas and wind facilities to describe the ongoing digital transformation as a process of infrastructuring in which the infrastructure gets increasingly entangled with internal and external systems, stakeholders, and agendas. In doing so, we shed light on how the established local and situated solutions evolve and are compensated for through the technical and organizational principles of the emerging information infrastructure.

Research methods are part and parcel of good research. Despite this, we often teach research methods in an ad hoc manner and do not reflect on relevant pedagogical aspects. This situation is changing due to the introduction of dedicated research methods courses as part of a transferable skills agenda in many universities. To address pedagogical aspects of teaching methods in IS, we organized a workshop during the 10th Scandinavian Conference on Information Systems (SCIS/IRIS). Our goal was to create a community of practice among supervisors and teachers of research methods in IS. In this paper, we present our findings from the workshop presentations and discussions. Three main themes emerge from our workshop: 1) Coping with student motivational aspects of learning research methods at micro and macro levels, 2) Challenges related to course design for different institutional contexts, and 3) The challenging interplay between IS theory and empirical data when teaching research methods. We discuss these themes and their relation to the field of IS. We propose some future research topics for the community.

Researchers of Information Systems have long demonstrated that the teaching of research methods hinges on the dynamic interplay between theory and practice. However, grasping the role of theory in both the process (research design) and products (reading and writing articles) of research is challenging for teachers and, crucially, their students. This project contributes to the pedagogical discourse on qualitative research methods in Information Systems. It proposes the prototype of a framework in the shape of an agile lecture sequence for making explicit the role of theory across the research process when teaching research methods in Information Systems. The proposed framework is based on a constructivist approach, recognizing that students co-construct their knowledge recursively through the interplay with the socio-cultural context. Copyright (c) 2020 Elena Parmiggiani Creative Commons License This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License.

Extant Information Systems research emphasizes the strategic benefits of digitalization and value co-creation for business. Less is known, however, about the dynamics of how value is co-created in the digitalization of the public sector, where data infrastructures are increasingly adopted. We identify three core empirical challenges for value co-creation in the public sphere, corresponding to the following conceptual tenets: participation in infrastructuring processes, data curation, and data protection. We propose to draw on the Participatory Design tradition that permeates the Information Systems field in Scandinavia to critically harness the political meaning of value co-creation. Drawing on a two-year project on the design of data infrastructures in three areas of the public domain (environmental monitoring, healthcare, and smart cities), we contribute to Information Systems by proposing a research agenda consisting of three future directions for critical studies of value co-creation in data infrastructures in the public sector.

Open Access. CC BY-NC-ND license (Attribution-NonCommercial-NoDerivatives.

This thesis investigates the development of ICT solutions for performing real-time subsea environmental monitoring during oil and gas offshore operations. The research is based on a three-year case study of an international oil and gas company headquartered in Norway. The thesis is specifically focused on aspects of integration: how new tools, systems, and approaches are developed, and how existing ones are adapted to fit the existing systems and practices of the oil and gas company. The thesis has three goals. First, it analyzes how real-time environmental monitoring emerges as a distributed, interacting, and interconnected sociotechnical network (information infrastructure). Second, it identifies how the company’s initiatives are currently changing the representations of subsea environmental risk in remote and previously inaccessible areas like the Arctic region. Third, it reflects on the research methodology to address such a spatially distributed and long-term setting. On the theoretical level this work aims to contribute to the field of Information Systems and, marginally, to Computer-supported cooperative work. It is also inspired by Science and Technology Studies. A theoretical framework is developed connecting contributions from these three streams of literature in order to characterize the co-evolution of environmental monitoring infrastructures and their phenomenon of interest, environmental risk. This evolving relationship shows that, on the one hand, the perception of environmental risk as a problem is affected by the integration mechanisms that generate new knowledge. On the other hand, new conceptions of environmental risk have consequences for the maintenance and upgrade of environmental monitoring infrastructures. I call this process infrastructuring. The thesis provides a vivid empirical illustration of how the relationship between environmental risk and integration strategies unfolds in offshore environmental monitoring. By merging the theoretical framework and the empirical case, I argue in favor of a flexible understanding of the development of complex systems in organizational settings. Studies of infrastructure should account for the existence of pragmatic approaches to face persistent uncertainty and the fluid relationship between situated usefulness and a cross-geographical dimension. In line with existing debates in the literature, I demonstrate that the traditionally bounded distinctions of system design/development/implementation and of users/designers must be overtaken when studying infrastructures. On the methodological level, this thesis therefore advocates a deeper empirical analysis of the unfolding of infrastructures in organizational realities.

Offentlige tjenestetilbud er i økende grad levert gjennom digitale agenter basert på kunstig intelligens, med chatboten som eksempel i front. Parallelt med inntoget av chatboter og deres forventninger i både privat og offentlig sektor, er det dokumentert flere tilfeller hvor brukeres forventninger ikke innfris. Fagområdet rundt konversasjonsagenter er preget av en økende mengde forskning med utgangspunkt i agenter, brukere og deres kontekst. Der hvor en stor del av eksisterende forskning har tatt utgangspunktet i selve konversasjonsagenten, fokuserer denne oppgaven på brukerens utgangspunkt. Sammenlignet med andre studier er oppgavens fokus på den offentlige konteksten viktig, i den grad den adresserer begrensningene ved å benytte et verdiperspektiv hentet rett fra en tradisjonell markedsbasert kontekst. Oppgavens formål tar utgangspunkt i brukeres verdiperspektiv for å forklare brukerinteraksjon og teknologiadopsjon ved AI-baserte konversasjonsagenter. Konkret spør oppgaven: «Hvordan påvirker design og verdioppfatning brukerinteraksjon med AI-baserte konversasjonsagenter i offentlig sektor?» Gjennom empiriske resultater fra oppgavens datainnsamling, basert på en spørreundersøkelse og 15 intervjuer, viser studiets resultater teknologiadopsjon sett fra et brukerperspektiv. Teknologiakseptansemodellen VAM (Value-based Adoption Model) forklarer funnene i studiet utfra et verdiperspektiv bestående av kostnads- og nyttefaktorer, og tolkes samtidig i lys av teori basert på offentlig verdi. Studien viser at forståelsen om brukeres verdioppfatning av chatboter i offentlig sektor ikke alene bør forklares utfra et tradisjonelt «effektivitets-basert» perspektiv, og at brukeres opplevelse av offentlig verdi kan forklare holdninger og oppførsel. Offentlig verdi er sterkt tilknyttet kvaliteten på den offentlige tjenesteleveransen, og fordelene som chatboter gir i denne sammenhengen samsvarer godt med dokumenterte faktorer. Der hvor brukerens formål samsvarer med chatbotens evne til å fremstå pålitelig og kompetent, eksisterer det en klar nytteverdi for tidsbesparelse og enkel tilgjengelighet sammenlignet med en menneskebasert tjenesteleveranse. Samtidig er det klart at sosial verdi i form av gjennomsiktighet er en betydelig faktor som påvirker brukerinteraksjon. Der hvor brukere mangler innsyn i prosessen rundt interaksjonen med chatboter, reduseres tillit og brukerens motivasjon til å benytte ikke-menneskelige kommunikasjonskanaler. I det hele er et offentlig verdiperspektiv nødvendig for at man skal få en mer helhetlig forståelse av hvilke verdier som ligger til grunne for brukeres teknologiadopsjon av AI-baserte konversasjonsagenter i offentlig sektor. Public services are increasingly delivered through digital agents based on artificial intelligence, with the chatbot as a primary example. At the same time as chatbots and their expectations rise, both in the public and private sector, examples of failures to meet said expectations raise along with them. Literature surrounding conversational agents have seen considerable growth, concerning areas connected with the users, the agents, and the context in which they are used. While the majority of the existing literature focuses on the conversational agent, the concern of this study focuses on the user-perspective. Especially significant when compared to similar studies is the focus on the public context and addresses the limitations of using a purely market-based understanding to comprehend said context. This study seeks to utilize user-perspective to explain user-interaction and technology-acceptance regarding AI-based conversational agents. The specific question being asked is: “How does design and perceived valued affect user-interaction of AI-based conversational agents in the public sector?”. The study’s empirical results, based on a survey and 15 interviews, shows technology-acceptance from a user perspective. The technology-acceptance model VAM (Value-based Adoption Model) explains the study’s findings from a value-perspective through benefits and sacrifices, and views this through the understanding of public value. This study shows that the understanding of users’ perceived value of chatbot in the public sector cannot alone be explained through a traditional “effectiveness-based” perspective, and that public value can be used to explain users’ attitude and behavior. Public value is strongly linked to the quality of the public service, and the advantages attributed to the chatbot in this context correlates well with previously documented factors. It exists a clear usefulness for time saving and easy accessibility, where users purpose correlates with chatbots ability to appear reliable and competent, compared to human-based service provision. At the same time is social value in shape of transparency clearly a significant factor that impact user interaction. Users’ motivation and trust to use non-human communication channels is reduced when the user lack transparency in the process regarding the interaction with chatbots. Overall is the public value perspective necessary to obtain a completely understanding of which values that is essential for users’ technology adoption of AI-based conversation agents in the public region.

Digitale plattformer finnes overalt rundt oss. De tilbyr tjenester som mange bruker daglig. Konkurranse, brukerbehov og et plattformøkosystem i stadig endring tydeligjør verdien av tjenesteinnovasjon. Det finnes et mangfold av forskjellige designrammeverk og teknikker som brukes for å utføre denne innovasjonen. Design science er nødvendig mer enn noen gang ettersom problemer i tjenesteinnovasjon utforskes i stadig nye miljøer. Å bruke velkjente og utprøvde metoder på eksisterende designrammeverk kan forbedre deres evne til å takle kompleksiteten til problemene som digitale plattformøkosystemer presenterer. Data science er et modent felt og tilbyr metoder som synergi godt med datagenereringen digitale plattformer kan tilby. Målet med dette studiet er å gi innsikt i hvordan digitale transaksjonsplattformer utfører tjenesteinnovasjon gjennom design science. Det praktiske målet med denne studien er å se etter sammenheng mellom brukervalg av påloggingsløsning og betalingsalternativ i en nettbutikk. Nettbutikken fungerer som designgjennstanden og målet for designrammeverket for design science. Data science testes som et verktøy for å utføre evaluering av designgjennstanden i designrammeverket ved å samle inn kvantitative data om bruken av komponentene i appen. I tillegg brukes kvalitative data fra spørreskjemaer til identifisere relevans og bekrefte funn av dataanalyse. Digital platforms are found everywhere around us. They provide services which many use on a daily basis. Competition, user needs and an ever-changing platform ecosystem makes the importance of service innovation abundantly clear. There exists a plethora of different design frameworks and techniques utilized to perform this innovation. Design science is needed more than ever as the problems of service innovation become more diverse and expand and create new environments. Applying rigorous methods to existing design frameworks can enhance their ability to tackle the complexity of the problems which digital platform ecosystems present. Data science is a mature field and provides rigorous methods which synergises well with the data generation digital platforms can provide. The aim of this study is to provide insight into how digital transaction platforms perform service innovation through design science. The practical goal of this study is to look for correlation and possible synergies between user choice of sign-in solution and payment option in an eCommerce store environment. The eCommerce store acts as the artifact and the target of the design science research framework. Data science is tested as a tool for performing evaluation of the artifact in the design science research framework by gathering quantitative data on the use of the app components. Supplemental qualitative data from questionnaires will identify relevance and confirm findings of data analysis.

This open access book presents a number of case studies on digital transformation in Norway, one of the fore-runners in the digital progress index established by the European Commission in 2020. They explore the process of adoption, diffusion and value generation from digital technologies, and how the use of different digital solutions has enabled Norwegian enterprises to digitally transform their operations and business models. The book starts with an introductory chapter summarizing a vast body of literature in order to synthesize what is already known about digital transformation before exploring the Norwegian context in more detail. Then a series of case studies from the private and public sector in Norway is presented. They document a process perspective which describes the sequence of events during and after adoption of digital solutions, as well as the types of business value that were realized. Through these single studies, the process of digital transformation is illustrated, a number of key findings highlighted, and eventually theoretical and practical recommendations based on these cases emphasized. The book closes with a brief overview of some emerging technologies, and comments on how they are likely to change different sectors. Digital transformation has been one of the priority areas for the Norwegian government over the past years and puts Norwegian enterprises upfront in adopting novel technologies and utilizing them for achieving organizational goals. This experience accumulated over the years makes the Norwegian context a particularly interesting one in understanding how private and public organizations make use of new digital solutions, what lessons can be learnt during the process, and what are some of the key success and failure factors. This way the book is written for practitioners who are currently involved in digital transformation projects in their organizations, researchers of information systems and management, as well as master students in degrees of informatics and technology management.

Miljødata blir brukt til å informere, begrunne, og evaluere miljøpolitikk. Dermed er det avgjørende at denne dataen både er av høy kvalitet, og at den er presis i sin gjenspeiling av den virkelige verden. Hvordan forskere oppnår denne kvaliteten på data har blitt undersøkt fra tekniske og praktiske perspektiv, med lite fokus på innvirkning fra forskere og deres interaksjon. Tidligere litteratur har identifisert utfordringer forskere opplever i å forberede og dele data, der fokuset er eksterne innflytelser slik som mangel på ressurser og utfordrende retningslinjer. Det er derimot lite fokus på interne prosesser og hvordan disse påvirker hverandre, samt hvilken rolle eksterne innflytelser har på forberedning og deling av data. Dette prosjektet består av en fortolkende case-studie. Prosjektet har som mål å bidra med ny empirisk innsikt i hvordan samarbeid og koordinativ praksis i dataforberedelse og datadeling foregår på norske miljøforskningsinstitutt, og hvordan denne praksisen påvirker datakvalitet. Case-studien er basert på kvalitative data i form av intervjuer, observasjoner, og dokumentasjon. Funnene i studien identifiserer viktige roller og avhengigheter i dataforberedelsesprosesser. Videre antyder funnene at dataforberedelse blir påvirket av flere typer avhengigheter, og at disse videre påvirker datakvalitet. Funnene viser også at forskere fra norske forskningsinstitusjoner ikke alltid opplever press i form av retningslinjer og annen politikk i like stor grad som tidligere forskning viser, som kan være tegn på at innføring og håndheving av disse varierer i stor grad, samt er avhengig av lokal politikk. Environmental data provides us with knowledge about the state of our environment, and is used as a foundation for informing, establishing, and revising environmental policy. Thus, the quality of this data, and the ability of this data to reflect the real world are crucial for policy decision-makers. Still, there seems to be a gap in existing literature when it comes to how this quality is achieved through relations researchers have with one another and their data, with most studies focusing on technical and practical measures. Previous literature has identified challenges in preparing and sharing environmental data, with the focus being on external influences such as resources or policy. With existing studies documenting these interrelationships, it seems fitting to look at this from a different perspective, exploring the internal processes of environmental data work, how they influence each other, and how external factors play into the preparation and sharing of data. This project consists of an explanatory case study, with an interpretive approach. The aim of the study is to contribute with empirical data on the coordinative and cooperative practices of researchers preparing and sharing data in Norwegian environmental research stations, and how these practices influence data quality. The case study draws on qualitative data in the form of interviews, observations and document studies. The findings identify key roles and dependencies in environmental data preparation. Furthermore, the findings suggest that data preparation is a cooperative activity that relies on multiple dependencies to function, and that the quality of the resulting data is a product of these dependencies. Additionally, the findings suggest that data sharing activities in Norwegian non-profit organizations are often much less pressured by regulatory measures in contrast to previous literature on the topic, suggesting that there is substantial variance in matters of local policy implementation.

Verifiable credentials, coupled with decentralized ledger technologies, have been potential providers of trustworthy digital identity for individuals, organizations, and other entities, and thus, potential enablers of trustful digital interactions. The rapid development of this technology—called self-sovereign identity (SSI)—and the ecosystems built around it have been fostered even more by the societal needs stemming from the current pandemic crisis, when governments, non-profit organizations, businesses, and individuals are working together on different aspects of SSI to enable mainstream adoption. In this study, we build on rich qualitative data gathered from SSI practitioners to give a fresh overview of the perceived benefits and challenges of SSI. The paper advances research on the domain of SSI adoption and provides valuable insights into the feasibility of SSI for practitioners both in the private and public sectors peerReviewed

Due to COVID-19, companies were forced to adopt new work processes, and reduce modern work environments such as collaboration spaces. Professionals from many fields were forced to work remotely, almost overnight. Little is known about the impact of such non-volunteer remote work on productivity, stress, and other key aspects of work performance. To further our understanding of the impacts of this situation and remote work in general, we conducted an exploratory study by studying 28 knowledge work professionals (researchers, software developers, interior designers, service designers, and development consultants) from the viewpoint of perceived productivity and aspects affecting it in this unusual setting. Early results showed the positive influence of self-efficacy and teamwork on productivity during remote work, while no moderating effect of measured physical stress on productivity either through the intrinsic or social factor was present. peerReviewed

Bruk av kunstig intelligens i offentlig tjenesteleveranse er stadig økende, og målet om å legge til rette for et digitalt førstevalg skal gjøre digital samhandling til normen for kontakt mellom innbygger og forvaltning. Flere offentlige aktører har tatt i bruk chatboter for innbyggerkontakt for å oppnå en høyere grad av selvbetjening, ressurssparing og økt tilgjengelighet. For å sikre utbredt akseptanse av chatboter er offentlig sektor avhengig av at brukere (innbyggere) har tillit til tjenesten. Tidligere forskning foreslår en rekke faktorer som påvirker utvikling av tillit i en menneske-chatbot interaksjon, deriblant oppfattet kompetanse. Mye av forskningen er derimot gjort på chatboter for kundeservice og i kommersiell sammenheng, og det er et gap i litteraturen om tillit til chatboter i offentlig sektor. Denne studien har som mål å bidra med ny empirisk innsikt om hvordan kompetanse påvirker brukeres tillit til chatboter med følgende forskningsspørsmål: Hvordan påvirker oppfattet kompetansenivå tillit mellom bruker og chatbot? For å besvare forskningsspørsmålet er det gjennomført en casestudie på bruk av chatbot i offentlig sektor, eksemplifisert med NAVs chatbot, Frida. Datainnsamlingen består av intervjuer gjennomført med informanter hos NAV samt en spørreundersøkelse besvart av innbyggere som har benyttet seg av chatbot for kontakt med offentlig sektor i Norge. Funnene viser at tillit i stor grad påvirkes av oppfattet kompetansenivå, og at brukeres forventninger før bruk spiller en viktig rolle i hvordan kompetanse oppfattes. Når brukere har høye forventninger til en chatbots kompetanse, innhold og evne til å forstå, og disse ikke innfris ved bruk, svekkes tilliten fort. Forventninger som i større grad stemmer overens med en chatbots reelle kompetansenivå fører til at tilliten forblir uendret ved bruk. Videre presenteres tiltak som kan legge til rette for tillit i en menneske-chatbot interaksjon. The use of artificial intelligence in public service delivery is constantly increasing, and the goal of promoting a ''digitalt førstevalg'', or a digital first choice, makes digital interaction the norm for interaction between citizens and public administration. Several public agencies are using chatbots for citizen contact to facilitate a higher degree of self-service, increase accessibility, and save resources. In order to achieve a widespread acceptance of chatbots, the public sector is dependent on users (citizens) trusting the service. Previous research suggests several factors that influence the development of trust in a human-chatbot interaction, including perceived competence. Although much research has been done on chatbots in recent years, most of it is done on chatbots for customer service and in a commercial context, and there is a gap in the literature on trust in chatbots in the public sector. This study aims to contribute new empirical insight into how competence affects users' trust in chatbots with the following research question: How does the perceived level of competence affect trust between user and chatbot? To answer the research question, a case study has been carried out on the use of chatbots in the public sector, exemplified by NAV's chatbot, Frida. The data collection consists of interviews conducted with informants from NAV as well as a survey answered by citizens who have used chatbots for contact with the public sector in Norway. The findings show that trust is largely influenced by perceived level of competence and that users' expectations before use play an important part in how competence is perceived. When users have high expectations of a chatbot's competence, content, and ability to understand, and these expectations are not met during the interaction, the users' trust quickly weakens. Expectations that correspond to a chatbot's real level of competence lead to their expectations being met during the interaction, and the trust remains unchanged. Furthermore, measures that can facilitate trust in a human-chatbot interaction are presented.

Bruk av AI-kontrollerte chatboter blir stadig vanligere både i privat og offentlig sektor. I Norge er adapsjonen innenfor offentlig sektor en del av en overordnet digitaliseringsstrategi som skal gi mer effektive og bedre tjenester til landets innbyggere. For at innbyggerne skal ta i bruk de nye digitale tjenestene som dukker opp som følge av digitaliseringen, er det viktig at tjenestene skaper og ivaretar tilliten hos befolkningen. Innbyggernes brukeropplevelse i møte med chatboter er viktige for å ivareta og styrke denne tilliten. Chatboter er i dag tatt i bruk av flere statlige og kommunale organisasjoner og fungerer som en del av deres førstelinje-kundestøtte. De er alltid tilgjengelige og har stor kapasitet til å svare på henvendelser fra innbyggerne. Dette frigjør arbeidstid fra mennesker som ellers måtte besvare de samme henvendelsene. Chatboter som brukes på denne måten er oppgave-orienterte og det er avgjørende for brukeropplevelse og tillit at de klarer å svare godt på brukernes henvendelser. Tidligere forskning har avslørt at enkelte grupper blant innbyggerne får problemer i møte med denne teknologien. Enkelte forstår ikke at de snakker med en AI-kontrollert chatbot, andre er ikke villige til å gi chatboten en sjanse og ber om å få snakke med menneskelige veiledere med gang. En tredje gruppe interagerer ikke med chatboten i det hele tatt etter å ha tatt kontakt. Studien tar utgangspunkt i eksisterende teori om tillit til AI generelt og AI-kontrollerte chatboter, både i en privat og offentlig setting, samt brukeropplevelser i møte med chatboter. Denne forskningen har som mål å bidra med empirisk innsikt i hvordan en statlig etat jobber med videreutvikling av chatbot-teknologien, med hensyn til tillit og brukeropplevelser og bygge videre på eksisterende teori innen feltet menneske-maskin interaksjon. Forskningsspørsmålet som besvares er: Hvordan påvirker tillit til et system brukeropplevelsen. Systemet det er snakk om er i denne sammenhengen er chatboten Frida, som representerer NAV. Forskningen er gjennomført som en deskriptiv casestudie, og har et kvalitativt forskningsdesign for å undersøke forholdet mellom tillit og brukeropplevelsen av chatboter innen offentlig sektor. Datainnsamling er gjort ved hjelp av semistrukturerte intervjuer med ansatte hos NAV som jobber med Frida. I tillegg er det gjennomført en spørreundersøkelse i mars-april for å samle erfaringer fra brukere av chatboter innen offentlig sektor. Dataanalysen er gjennomført i henhold til en stedvis-deduktiv metode. Funnene fra undersøkelsen beskriver hvordan NAV jobber med videreutvikling av chatboten sin, hvilke tanker og vurderinger de gjør med tanke på å legge til rette for en god brukeropplevelse og skape tillit blant brukerne. Svar fra spørreundersøkelsen viser respondentens holdninger og erfaringer med digitalisering og automatisering av offentlige tjenester. Det avslører et bredt spekter av forventninger, og hvordan disse påvirker brukeropplevelsen i de individuelle tilfellene. Dette danner et grunnlag for om tilliten til systemet styrkes eller svekkes i etterkant av møte med chatboter. The use of AI-controlled chatbots is expanding both in the private and public sectors. In Norway, the adaption of these types of systems is part of an overarching strategy for digitalization of the public sector. One of the goals of this ongoing process is to provide better and more efficient services to the public at large. It is important to create and maintain public trust to ensure that new services that appear because of the ongoing digitalization is adopted. The public's user experiences from meetings with chatbots is important to maintain and strengthen this trust. Chatbots are adopted by several public organizations and work as part of their first line of customer support. They are always available and have unlimited capacity to answer inquiries from the public. This frees up available workhours from human employees that would otherwise be needed to answer public inquiries. Chatbots that are used in this manner are task-oriented and it is crucial for the user experience they provide that they respond well to these inquiries. Prior research has revealed that parts of the public struggle when they encounter AIcontrolled chatbots. Some do not understand that they are talking to an AI. Others are not willing to interact with the chatbot and request a human councilor immediately. A third group contact the chatbot without engaging with it at all. This study is based on prior research into general trust in AI and trust in AI-controlled chatbots, both in a private and public setting, as well as research into user experiences regarding chatbots. This thesis aims to contribute with empiric insight into how a government agency work with continual development of chatbot technology regarding trust and user experiences, and to further develop existing theory in the field of human-machine interaction. To do this, the following research question is explored and answered: How does trust in a system affect user experience? The system in this case refers to Frida, a chatbot employed by NAV, the government agency running Norway's welfare programs. The research is done as a descriptive case study and employs a qualitative research strategy to examine the relationship between trust and user experience of chatbots in the public sector. Data was gathered through interviews with employees that work with Frida at NAV. Additionally, an online questionnaire was run in March-April to gather experiences from users of public sector chatbots. The data was analyzed using a stepwise deductiveinductive method. Findings from the interviews describe how NAV work with continual development of their chatbot, their thoughts and judgements regarding creation of a good user experience and creating and maintaining trust among the public. Findings from the questionnaire show attitudes and user experiences with digitization and automation of public services. It reveals a wide range of user expectations and show how these affect user experience in individual cases. This forms a basis for whether trust in the system is strengthened or weakened after a user's interaction with a chatbot.

Information systems (IS) implementation often aims to ensuring user satisfaction. However, achieving such user-centredness has remained ambiguous and challenging, and the results are not always those that were promised. This may result from several views and fluctuating and implicitly defined concepts. While some premises have been identified, they seem to mostly concern easily manageable settings where the number of users is limited, or where the possibility to tailor the system is significant. Especially in a large-scale system's implementation user-centredness seems to be fuzzy. In this paper we illustrate how user-centredness unfolds in a large scale IS implementation. We conduct a qualitative case study to see what occurs when the efforts are declared user-centred. By interviewing 13 central actors from a local developer organization, we learnt that user-centredness in such context is essentially the result of joint efforts thus necessitating that each party carries out their responsibility for user-centredness and engages in collaboration with others. The paper contributes to research by sharing empirically grounded findings to be used to extend the discussion on user-centredness. publishedVersion

Bruken av intelligente teknologier, IoT enheter, i forskning har bidratt til å gjøre arbeid med miljøforskning lettere og mer tilgjengelig, da det er mulig å måle nesten det en vil til en lav pris. Eksempler på intelligente teknologier er sensorer som måler temperatur, fuktighet eller trykk, kameraer og akustiske enheter slik som ekkolodd, og de er intelligente fordi de kan levere data automatisk uten behov menneskelig interaksjon. Automatiske innsamlinger av sanntidsdata, ofte fra enheter med ulik kvalitet og følsomhet, fører imidlertid til lagring av omfattende mengder med rådata. Som et resultat av dette må forskere kontinuerlig forsikre seg om at enhetene er kalibrerte, og gjøre dataene pålitelige, lesbare og meningsfulle med hensyn til konteksten. For å gjøre forskning gjennomførbart, er forskningsstasjoner – som også kan bli referert til som forskningsinfrastrukturer – noen ganger avhengige av finansiering via styrende institusjoner som for eksempel ESFRI og Forskningsrådet i Norge. For å motta finansiering – og med et økt fokus på åpen datadeling mellom forskningsinfrastrukturer, stilles det høyere krav til datadokumentasjon i henhold til de standarder og krav som myndighetene setter. Dette er med på å påvirke forskeres datastyringsarbeid. Forskningen som er gjort i dette prosjektet har som mål å bidra med empirisk innsikt i hvordan intelligente teknolgier brukes på forskningsstasjoner for miljøovervåkning, og hvordan forskeres arbeid påvirkes av økt bruk av teknologi i overvåkning. Det er undersøkt hvordan forskere opplever retningslinjene satt av forskningsinstitusjoner om hvordan dataene skal dokumenteres, og hvordan deres dataarbeid må tilpasses deretter. Det er gjennomført et casestudie av utvalgte forskningsstasjoner for miljøovervåkning i Norge, og studien baserer seg på kvalitative data fra strukturerte og semi-strukturerte intervjuer og relevante dokumenter fra myndighetene slik som strategidokumenter, retningslinjer og veikart. Funnene er basert på informasjon fra intervjuer av informanter som til daglig arbeider på forskningsstasjoner for miljøovervåkning som dataledere, forskningssjefer, forskningskoordinatorer eller miljøforskere. Funnene viser at IoT basert miljøovervåkning er muliggjort av datastyringsarbeid, med etablerte prosesser for å sikre at store sett av rådata blir gjort pålitelige, lesbare og meningsfulle for å støtte fremtidig gjenbruk og tolkning. Prosessene er påvirket av retningslinjer, prosedyrer og standarder for hvordan dataene skal samles inn og håndteres, som er med på å påvirke forskeres arbeidspraksiser. På grunn av begrensninger i tid, finansiering og ressurser, viser funnene imidlertid at det kan være vanskelig å ha ønsket kvalitet i forskernes vitenskapelige arbeid. The use of intelligent technologies, particularly IoT devices, in research has contributed to more accessible and available monitoring of the environment with the possibility of monitoring almost anything with low costs. Examples of intelligent technologies used in environmental research are sensors monitoring temperature, humidity or pressure, cameras, and acoustic devices like echo sounders. The devices are intelligent because they can automatically deliver data without human interaction. However, automatic collections of real-time data from sensor devices with heterogeneous quality and sensitivity lead to extensive raw data storage. Subsequently, researchers must ensure that the devices are calibrated and ensure that the data are trustworthy, readable, and meaningful to the context. To make research feasible, research institutions – which can also be referred to as research infrastructures – sometimes are dependent on receiving funding from governing institutions such as ESFRI and the Research Council of Norway. In order to receive funding – and with an increased focus on open data sharing between research infrastructures, researchers must adapt to the governments' guidelines and requirements to document their research, influencing their data governance activities. This research aims to contribute empirical insights into the use of intelligent technologies at research infrastructures for environmental monitoring and how researchers' work is affected by the increased use of technology in monitoring. Additionally, researchers' experience adapting to guidelines and requirements on data documenting is also investigated. It is conducted a case study of selected research infrastructures for environmental research in Norway. The study is based on qualitative data collected from structured and semi-structured interviews and relevant documents from governing institutions such as strategy documents, guidelines, and roadmaps. The findings are based on information retrieved from interviews of data managers, environmental researchers, research coordinators, and research managers that daily work at research infrastructures for environmental monitoring. The findings show that IoT-based environmental monitoring is enabled by data governance with the established processes to translate raw, often big data sets, into reliable, readable, and meaningful information to support future reuse and interpretation. The processes are affected by policies, procedures, and standards on collecting and managing the data that consequently affect researchers’ work practices. However, due to constraints in time, funding, and resources, the findings also show that it can be challenging to have the desired quality in researchers’ scientific work.

Formålet med denne oppgaven er å kunne se på hvordan data governance utspiller seg i praksis i en IT outsourcing. Ved å opparbeide seg en dypere forståelse for hva som skjer i en outsourcing fra outsourcing-selskapet sitt perspektiv, så har forskeren nøye forklart hvordan de relevante fasene av en outsourcing fungerer. Måten forskningen ble gjennomført på er basert på litteraturgjennomgang og et casestudie hos en verdensledende selskap innen IT outsourcing. Det har blitt gjort dybdeintervju med dette selskapet sine ansatte stasjonert i deres outsourcing-avdelinger. Resultatene fra prosjektet er et konseptuelt forslag til en modell som kan brukes for data governance i en outsourcing, basert på både formelle kontraktuelle krav og uformelle interpersonlige forhold. Det har kommet frem at disse to faktorene har en stor innflytelse på data governance i praksis, og disse har blitt tatt mye hensyn til i det konseptuelle forslaget til løsning. The purpose of this master's thesis is to be able to see how data governance in practice takes place in an IT outsourcing. The researcher will explain how the phases in an outsourcing work and the different phases it consists of, based on obtaining a deeper understanding of what is happening in an outsourcing from the perspective of the outsourcing-company. To examine this, the type of method chosen is a case study with a global market-leading company in IT outsourcing together with litterature review. This has been done through qualitative interviews with employees stationed in the companys outsourcing-department. The results of this research is a conceptual suggestion to a modell that could be used for data governance in an outsourcing, based on formal contractual demands and agreements, as well as the informal interpersonal relationship between the outsourcing company and the client. The research shows that these two elements have a big influence on data governance in practice, and these has been taken highly in to account in the conceptual suggestion for a solution.

Forskningsinfrastrukturer for miljøovervåking har blitt mer regulert av styrende organer, som ESFRI, gjennom retningslinjer og krav for å motta finansiering. Disse retningslinjene har hovedsakelig et teknisk fokus, mens de komplekse sosio-tekniske relasjonene mellom forskerne og den tekniske infrastrukturen ved forskningsinfrastrukturer ofte blir neglisjert. Tidligere forskning viser at disse retningslinjene skaper spenninger for forskerne, ettersom de kontinuerlig må balansere ulike interesser fra forskjellige aktører i det daglige arbeidet. I tillegg, er arbeidsmetodene for datainnsamling under kontinuerlig utvikling. Utvikling av ny teknologi har ført til at arbeidet med datainnsamling har blitt mer sensorbasert og automatisert. Tidligere forskningsartikler tar for seg den økte betydningen av data curation, som følge av at forskningsinfrastrukturene blir mer tekniske og automatiserte. Denne studien har som mål å bidra med ny empirisk innsikt om hvordan forskerne blir påvirket av å ta i bruk nye løsninger, ved å se på bruken av IoT-teknologi ved forskjellige forskningsinfrastrukturer, og hvordan det påvirker data curation, det økte fokuset på datadeling, og det daglige arbeidet til forskerne. Denne oppgaven tar i bruk et information infrastructure-perspektiv for å se på forskningsinfrastrukturer som komplekse sosio-tekniske infrastrukturer som er under kontinuerlig utvikling. Denne casestudien er basert på kvalitative data: Intervjuer, observasjoner under seminarer og dokumentasjon for å svare på forskningsspørsmålene, og inkluderte intervjuer fra miljøforskere og andre aktuelle informanter som jobber ved forskningsorganisasjoner for miljøforskning. Funnene beskrev hvordan data curation kunne forbedre bruken av IoT, støtte distribuert datadeling og viktigheten av data curation for å sikre at vitenskapelige data er pålitelige. Videre identifiserte studien ekstraarbeidet som kreves fra forskere, og hvordan de påvirkes av manglende insentiver, tid og ressurser. Research infrastructures for environmental monitoring has become more regulated by governing bodies, such as ESFRI, through policies and requirements to receive financing. These policies have a technical focus primarily, while the complex socio-technical relations present at research infrastructures often get neglected. Previous research shows that these policies cause tension for the researchers, as they have to continually balance the interests of different actors in their daily work. Furthermore, data collection methods evolve with the development of new technologies, becoming more sensor-based and automated. Previous articles address the increased importance of data curation, as the research infrastructures become more technical and automated. This study aims to contribute with new empirical insight into how the researchers are affected by the adoption of new technologies, by looking at the use of IoT technology at different research infrastructures, and how it impacts the data curation, the increased focus on data sharing, and the researchers' daily work. This thesis adopts an information infrastructure perspective to view research infrastructures as continuously evolving and complex socio-technical infrastructure. This interpretive case study relied on qualitative data: Interviews, observations during seminars, and documentation to answer the research questions, and included environmental researchers and other participants of interest working at environmental research organizations. The findings describe how data curation could improve the use of IoT, support distributed data sharing, and the importance of data curation to ensure that scientific data is trustworthy. Furthermore, this thesis identifies the extra work required by the researchers, and how they are affected by the lack of incentives, time, and resources.

Industirer I dag blir møtt av et fort marked for digitale tjenester tilbudt av IT selskaper, og gjennom den digitale æra utvalget av tjenester har økt som følger av konstant innovasjon. Men ettersom den integrerte teknologien blir mer avansert, så gjør også kompleksiteten i de utviklede infrastrukturene for informasjon. Målet med denne avhandlingen er å undersøke outsourcing av tjenester som fenomen og skape rik innsikt i den praktiske utviklingen av infrastrukturer for informasjon. Dette blir gjort ved en sosioteknisk tilnærming, ved undersøkelse av tekniske og sosiale faktorer, og deres interaksjoner, i det beskrevede fenomenet. For å bidra med denne kunnskapen en case studie har blitt gjennomført for å undersøke prosesser, elementer, aktører, interaksjoner og utfordringer som tar plass gjennom outsourcing av tjenester, tilbudt av et av de globalt ledende IT selskapene. Gjennom semi-strukturerte intervjuer og en stegvis-deduktiv induserende metode for data analyse, konsepter har blitt manet frem, karakterisert og validert i litteratur fra forskningsområdet for informasjonssystemer i tillegg til gjennom diskusjon med informantene. Disse karakteristikkene har videre blitt modellert og som resultat blitt til et rammeverk for outsourcing a v tjenester som illustrerer kompleksiteten av fenomenet. Resultatene indikerer at de identifiserte faktorene for kompleksitet er gjensidig avhengige og paradoksale. Men en potensiell mulig løsning, en minimum data modell for brukbarhet har dukket opp og blir presentert som motivasjon for videre forskning. Industries today are faced with a large market of digital services offered by IT corporations, and through the digital age the array of services has increased though constant innovation. However, with the integrated technologies advancing the complexity of the information infrastructures developed increases. The aim of this thesis is to investigate the phenomenon of digital service outsourcing in order to create new insight into the practical development of Information Infrastructures. This is done by adapting a sociotechnical approach, studying the technical and social factors, as well as their interactions, in the described phenomenon. In order to provide this knowledge a case study has been conducted, investigating the processes, elements, actors, interactions and challenges occurring though service outsourcing delivered by an IT corporation among the global leaders in the market. Through semi-structured interviews and a stepwise-deductive inductive data analysis emerging concepts have been characterized and further validated in Information System literature and though discussion with the informants. These characteristics have been modeled, resulting in a service outsourcing framework, illustrating the complexity of the phenomenon. Further the results indicate that the complex factors identified are interdependent and paradoxical. However, an emerging potential solution of a minimum viable data model has also been identified and is presented as motivation for further research.