Denna studie undersöker de retoriska strategierna som används i debattartiklar publicerade av den svenska tidningen Svenska Dagbladet angående användningen av ChatGPT, en avancerad chatbot, inom utbildning. Analyserna av dessa artiklar syftar till att belysa argumenten för och emot ChatGPT:s användning i skolor och bidra till en bättre förståelse av hur retorik kan forma åsikter om teknologi och innovation inom utbildning. För att uppnå detta använder studien en teoretisk ram som hämtar från konstruktivismteori, medialisering, moralpanik, teknisk determinism och retorikteori och dess övertygande enheter. Forskningsfrågorna som utforskas i denna studie inkluderar skildringen av ChatGPT:s påverkan på utbildning, de retoriska strategierna som används i artiklarna och den potentiella påverkan av moralisk panik på attityder till teknologi och innovation inom utbildning. Resultaten avslöjar att artiklarna presenterade både positiva och negativa perspektiv på användningen av ChatGPT och att olika retoriska strategier, såsom appel till auktoritet och känslor, användes för att påverka läsarnas åsikter. Studien bidrar till en bättre förståelse av retorikens roll i att forma attityder till användningen av avancerade chatbots inom utbildning och betonar vikten av att överväga flera teoretiska perspektiv vid analys av mediediskurs. / This study examines the rhetorical strategies used in opinion articles published by the Swedish newspaper Svenska Dagbladet regarding the use of ChatGPT, an advanced chatbot, in education. The analysis of these articles aims to shed light on the arguments for and against ChatGPT's use in schools and to contribute to a better understanding of how rhetoric can shape opinions about technology and innovation in education. To accomplish this, the study employs a theoretical framework that draws from constructivism theory, mediatization, moral panic, technological determinism, and rhetoric theory and its persuasive devices. The research questions explored in this study include the portrayal of ChatGPT's impact on education, the rhetorical strategies used in the articles, and the potential influence of moral panic on attitudes towards technology and innovation in education. The findings reveal that the articles presented both positive and negative perspectives on the use of ChatGPT, and that various rhetorical strategies, such as appeals to authority and emotion, were used to influence readers' opinions. The study contributes to a better understanding of the role of rhetoric in shaping attitudes towards the use of advanced chatbots in education and highlights the importance of considering multiple theoretical perspectives in analyzing media discourse.

Digitalisering är under ständig utveckling inom banksektorn. I och med den växande utvecklingen så har intresset för artificiell intelligens ökat men trots detta finns det fortfarande kvar osäkerheter kring användandet av AI-chatbots. Banker hanterar en stor mängd känsliga uppgifter vilket gör att konsumenters upplevda oro och tillit är extra känsligt. Detta är viktigt att understryka då tidigare forskning har visat att tillit utgör en stor del av konsumenters beslut att använda AI-chatbot. Baserat på tidigare forskning har denna studie knutit samman teorier om vilka faktorer som påverkar svenska bankkonsumenters tillit till AI chatbots samt vilka faktorer som påverkar svenska bankkonsumenters intention att använda AI chatbots vid utförande av bankärenden online. Detta utgör även studiens syfte. Det teoretiska ramverket i studien utgörs av technology acceptance model (TAM), on-line trust-modellen och upplevd mänsklighet. Utifrån dessa teorier har studiens hypoteser byggts upp där vardera hypotes testar en faktor som beskrivs i de valda teorierna. Datainsamlingen skedde genom en enkätundersökning online som distribuerades på sociala kanaler. Undersökningen genererade 87stycken fullständiga svar. Insamlade data analyserades med hjälp av Minitab och Excel. De analysmetoder som använts för att analysera insamlade data var huvudsakligen korrelationstest och regressionsanalys. Resultatet visade att den upplevda nyttan och tillit är faktorer som påverkar intentionen till användning positivt. Resultatet visade även att faktorerna trovärdighet och användarvänlighet påverkar tilliten positivt samt att risk har en negativ påverkan på tilliten. För att effektivisera användning av AI-chatbots som kundtjänstverktyg så rekommenderas svenska banker att fokusera på att öka den upplevda nyttan i form av att säkerställa snabbhet, produktivitet och enkelhet. Svenska banker bör även säkerställa att bankkonsumenter känner tillit till AIchatbots. Detta kan göras genom att fokusera på ökad användarvänlighet och trovärdighet samtidigt som banker bör minska den upplevda risken i form av integritetsrisk då detta är faktorer som visat samband med tillit till AI-chatbots inom banksektorn. / Digitalization is under constant development within the banking industry. With this comes a growing interest in artificial intelligence. Despite this, uncertainty regarding the use of AI-chatbots remains. Banks handle a large amount of sensitive information which, for the consumers, makes the perceived trust especially delicate. Earlier research has shown that trust has a significant impact on the consumers intention to use AI-chatbots. Based on previous research this study aims to tie together theories about which factors impact Swedish bank customers’ trust towards AI-chatbots, as well as the customers’ intention to use AI-chatbots while performing bank-related business online. This is also considered to be the purpose of this paper. In this study the theoretical standpoint is based on the technology acceptance model (TAM), on-line trust and perceived humanness. From these well-established theories a set of hypotheses has been presented, where each hypothesis examines different factors that are presented in the used model. The data collection in this study was conducted through an online survey, where a total of 87 answers were collected. The data was analyzed using Minitab. The main analysis methods used were correlation tests as well as a regression analysis. The results showed that the perceived usefulness and trust are factors that have a positive impact on the intention to use AI-chatbots while performing banking tasks. The results also showed that ease of use and credibility have a positive impact on trust while risk has a negative impact on trust toward AI-chatbots. To make the use of AI-chatbots more effective as a customer service tool this study recommends Swedish banks to focus on increasing the perceived usefulness of the AI-chatbots. This can be done by ensuring the AI-chatbots are able to provide productivity, simplicity and swiftness to its users. Swedish banks should also focus on increasing customers’ perceived trust in the AIchatbots. This can be achieved by increasing the ease of use as well as the credibility of the system. To decrease the experienced risk of using the AI-chatbots, this study recommends banks to focus on decreasing the integrity risk.

Embodied Conversational Agents (ECA) form part of a range of virtual characters whose intended purpose include engaging in natural conversations with human users. While works in literature are ripe with descriptions of attempts at producing viable ECA architectures, few authors have addressed the role of episodic memory models in conversational agents. This form of memory, which provides a sense of autobiographic record-keeping in humans, has only recently been peripherally integrated into dialog management tools for ECAs. In our work, we propose to take a closer look at the shared characteristics of episodic memory models in recent examples from the field. Additionally, we propose several enhancements to these existing models through a unified episodic memory model for ECA's. As part of our research into episodic memory models, we present a process for determining the prevalent contexts in the conversations obtained from the aforementioned interactions. The process presented demonstrates the use of statistical and machine learning services, as well as Natural Language Processing techniques to extract relevant snippets from conversations. Finally, mechanisms to store, retrieve, and recall episodes from previous conversations are discussed. A primary contribution of this research is in the context of contemporary memory models for conversational agents and cognitive architectures. To the best of our knowledge, this is the first attempt at providing a comparative summary of existing works. As implementations of ECAs become more complex and encompass more realistic conversation engines, we expect that episodic memory models will continue to evolve and further enhance the naturalness of conversations.

Chattbotutvecklare på Softronic använder i dagsläget Rasa-ramverket och dess standardkomponenter för bearbetning av användarinmatning. Det här är problematiskt då standardkomponenterna inte är optimerade för det svenska språket. Till följd av detta efterfrågades en utvärdering av samtliga Rasa-komponenter med syfte att identifiera de mest gynnsamma komponenterna för att maximera klassificeringsträffsäkerhet. I detta examensarbete framtogs och jämfördes flera Rasa-pipelines med olika komponenter för tokenisering, känneteckensextrahering och klassificering. Resultaten av komponenterna för tokenisering visade att Rasas WhitespaceTokenizer överträffade både SpacyTokenizer och StanzaTokenizer. För känneteckensextrahering var CountVectorsFeaturizer, LanguageModelFeaturizer (med LaBSE-modellen) och FastTextFeaturizer (med den officiella fastText-modellen tränad på svenska Wikipedia) de mest optimala komponenterna. Den klassificerare som i allmänhet presterade bäst var DIETClassifier, men det fanns flera tillfällen där SklearnIntentClassifier överträffade den.   Detta arbete resulterade i flera pipelines som överträffade Rasas standard-pipeline. Av dessa pipelines var det två som presterade bäst. Den första pipeline implementerade komponenterna WhitespaceTokenizer, CountVectorsFeaturizer, FastTextFeaturizer (med den officiella fastText-modellen tränad på svenska Wikipedia) och DIETClassifier med en klassificeringsträffsäkerhet på 91% (F1-score). Den andra pipeline implementerade komponenterna WhitespaceTokenizer, LanguageModelFeaturizer (med LaBSE-modellen) och SklearnIntentClassifier med en klassificeringsträffsäkerhet på 91,5% (F1-score). / Chatbot developers at Softronic currently use the Rasa framework and its default components for processing user input. This is problematic as the default components are not optimized for the Swedish language. Following this an evaluation of all Rasa components was requested with the purpose of identifying the most favorable components to maximize classification accuracy. In this thesis, several Rasa pipelines were developed and compared with different components for tokenization, feature extraction and classification. The results of the tokenization components showed that Rasa's WhitespaceTokenizer surpassed both SpacyTokenizer and StanzaTokenizer. For feature extraction, CountVectorsFeaturizer, LanguageModelFeaturizer (with the LaBSE model) and FastTextFeaturizer (with the official fastText model trained on Swedish Wikipedia) were the most optimal components. The classifier that generally performed best was DIETClassifier, but there were several occasions where SklearnIntentClassifier surpassed it. This work resulted in several pipelines that exceeded Rasa’s standard pipeline. Of these pipelines, two performed best. The first pipeline implemented the components WhitespaceTokenizer, CountVectorsFeaturizer, FastTextFeaturizer (with the official fastText model trained on Swedish Wikipedia) and DIETClassifier with a classification accuracy of 91% (F1 score). The other pipeline implemented the components WhitespaceTokenizer, LanguageModelFeaturizer (with the LaBSE model) and SklearnIntentClassifier with a classification accuracy of 91.5% (F1 score).

Interaction with digital products and interfaces concern more and more of human decision-making and the problems regarding environmental, financial and social sustainability are consequences much due to our behavior. The issues and goals of sustainable development therefore implies how we have to think differently about digital design. In this paper, we examine the adequacy of influencing sustainable behavior with a data-driven design approach, applying a conversational user interface. A case study regarding the United Nation’s goals of technological development and economic distribution was conducted, to see if a hypothetical business with a proof-of-concept digital product could be effective in influencing where companies base their operations. The test results showed a lack of usability and influence, but still suggested a potential with language-based interfaces. Even though the results could not prove anything, we argue that leveraging data analysis to design for sustainable behavior could be a very valuable strategy. A data-driven approach could enable ambitions of profit and user experience to coincide with those of sustainability, within a business organization.

The recent release of AI tools for public use allows for the development of novel teaching approaches for goals that often present challenges in the classroom, such as the need for personalized learning materials. The current study enlists a four-week ChatGPT-based personalized learning intervention in tandem with a teacher questionnaire and interviews in two upper-secondary schools in Southern Sweden to investigate English teacher perceptions of the benefits and challenges of using AI for personalized language learning. In addition, the intervention investigates the potential effectiveness of personalized learning assignments using ChatGPT on the development of students’ grammar abilities in a specific, local classroom context to both address a local need at the school in question and to serve as a proof of concept for more broad-based, future research on the use of these tools for this purpose. The questionnaire revealed that teachers initially had some concerns regarding the accuracy, reliability, and practical implementation of such tools. However, the intervention was found to significantly reduce grammar errors in student writing, and in follow-up interviews, teachers reported feeling more receptive to such approaches after interacting with the tools and seeing the beneficial results. These findings demonstrate that teachers may be hesitant to implement AI tools, which underscores the importance of training and first-hand use for promoting their successful adoption into pedagogical practices. In addition, the findings suggest that AI-based tools for personalized language learning may also be successful in a broader educational context. Finally, certain limitations, such as the small sample size, are acknowledged which emphasizes that further research is necessary to acquire a more comprehensive understanding of personalized learning using AI-based tools like ChatGPT.

Intelligent conversational agents (ICAs) are artificial intelligence (AI)-enabled systems that can communicate with humans through text or voice using natural language. The first ICA, “Eliza,” appeared in 1966 to simulate human conversation using pattern matching. Commercial ICAs appeared on the AOL and MSN platforms in 2001 and aided in developing advanced AI and Human-Computer Interaction (HCI). Since then, ICAs have progressively appeared in consumer products and services. Their success depends on the user’s experience and attitude towards these services. This research examines customer attitudes towards ICAs through a theoretical framework of integrated Expectation Confirmation Theory (ECT) and Task Technology Fit Theory (TTF). By exploring user experience via an experiment that engages end-users with ICA’s different functions and tasks, this study examines user perception of ICA’s AI capabilities, such as Conversation Ability, Friendliness, Intelligence, Responsiveness, Task Performance, and Trust. This research investigates how customer satisfaction with ICA capabilities and perceived task technology fit influence their intention to use ICAs. A field survey of 380 Canadian end-users utilizing ICAs on the websites of five large Canadian telecom service providers enabled empirical testing of the model. / Thesis / Doctor of Philosophy (PhD)

ChatGPT är inte en vanlig AI-chatbot; det är ett kraftfullt verktyg som har tränats för att skapa naturligt språk och kan tillämpas inom högre utbildning. Genom sin förmåga att kunna generera text och svara på frågor kan det vara en ovärderlig tillgång för lärare. Att artificiell intelligens påverkar den högre utbildning kan innebära problematik, eftersom det kräver anpassning av lärares användning av teknik i sitt arbete. AI-chatbots som ChatGPT har potentialen att förbättra undervisningen genom att stödja lärare. Trots fördelar medför ChatGPT även utmaningar, inklusive möjligheten till felaktiga svar, vilket kräver en kontroll av kvalitet. Problematiken med fördomar i AI-verktyg kan resultera i orättvisa bedömningar av studenter. För att hålla jämna steg med den teknologiska utvecklingen måste lärare utveckla nya kompetenser. Syftet med den kvalitativa studien är att djupgående undersöka vilka möjligheter och utmaningar som uppstår när lärare inom högre utbildning använder ChatGPT i undervisningsarbetet. Upptäckterna från studien visar att användningen av ChatGPT inom högre utbildning öppnar upp både möjligheter och utmaningar för lärare. När det kommer till möjligheterna är det ett verktyg som är tids- och energisparande främst inom användningsområdena, planering och förberedelse av undervisningen. Dessutom främjas analys, jämförelse och kritiska tänkande hos studenter samt möjliggör det för skapandet av unika uppgifter och automatiserad bedömning inom examination och bedömning. Trots dessa fördelar framkom det också utmaningar som är kopplade till användningen av ChatGPT inom högre utbildning. Lärare står inför juridiska och etiska dilemman när det gäller verktygets användning inom examination och bedömning. Det finns också en risk för att generera inkorrekt information och bristande kompetens. För att hantera dessa utmaningar är det avgörande att lärare behåller sin expertis och prioriterar mänsklig interaktion för att skapa en engagerande och meningsfull lärandemiljö. Genom att hitta en balans mellan teknik och pedagogik kan lärare säkerställa att ChatGPT blir ett kraftfullt hjälpmedel i deras undervisningsstrategier. / ChatGPT is not just an ordinary AI chatbot; it is a powerful tool that has been trained to generate natural language and can be used in higher education. By being able to generate text and provide answers to questions, it can be an invaluable asset for teachers. The impact of artificial intelligence on higher education can pose challenges, as it requires adaptation of how instructors use technology in their work. AI chatbots like ChatGPT have the potential to enhance education by supporting teachers. Despite advantages, ChatGPT also presents challenges, including the possibility of inaccurate responses, necessitating quality control. The issue of biases in AI tools can result in unfair student evaluations. To keep pace with technological advancements, teachers must develop new competencies. The purpose of the qualitative study is to deeply investigate the opportunities and challenges that arise when educators in higher education use ChatGPT in their teaching practices. The purpose of this qualitative study is to thoroughly examine the use of ChatGPT as a tool for teachers in higher education to facilitate their teaching work. The findings from the study reveal that the use of ChatGPT in higher education presents both opportunities and challenges for teachers. In terms of opportunities, it is a time- and energy- saving tool, particularly in the areas of lesson planning and preparation. Additionally, it promotes analysis, comparison, and critical thinking among students, and enables the creation of unique assignments and automated assessment in examination and evaluation. Despite these advantages, there are also challenges associated with the use of ChatGPT in higher education. Teachers face legal and ethical dilemmas when it comes to utilizing the tool in examination and assessment. There is also a risk of generating inaccurate information and lacking expertise. To address these challenges, it is crucial for teachers to maintain their expertise and prioritize human interaction to create an engaging and meaningful learning environment. By finding a balance between technology and pedagogy, teachers can ensure that ChatGPT becomes a powerful tool in their teaching strategies. This study is written in the Swedish language

Tesis por compendio / [ES] El mundo es testigo de nuevos avances tecnológicos que afectan significativamente a las organizaciones en diferentes departamentos. La inteligencia artificial (IA) es uno de estos avances, visto como una tecnología revolucionaria en la gestión de recursos humanos (RRHH). Profesionales y académicos han discutido el brillante papel de la IA en RRHH. Sin embargo, el análisis profundo de esta tecnología en el proceso de RRHH es aún escaso. Con todo ello, el objetivo principal de esta tesis es investigar el estado de la IA en RRHH y así identificar factores clave de implementación concretos. Primero, construyendo un marco académico para la IA en RRHH; segundo, analizar las aplicaciones de IA más utilizada en los procesos de RRHH; tercero, identificar las formas óptimas de transferir el conocimiento en los procesos de implementación de IA. La metodología utilizada para la investigación combina la revisión sistemática de la literatura y técnicas de investigación cualitativa. Como base y medida preparatoria para abordar las preguntas de investigación, se llevó a cabo un extenso análisis de la literatura en el campo AI-RRHH, con un enfoque particular en las publicaciones de algoritmos de IA en HRM, análisis de HR-Big data, aplicaciones/soluciones de IA en HRM e implementación de IA. En la misma línea, el autor publicó artículos en varias conferencias que contribuyeron a mejorar la madurez de las preguntas de investigación. Con base en este conocimiento, los estudios publicados ilustraron la brecha entre la promesa y la realidad de la IA en RRHH, teniendo en cuenta los requisitos técnicos de la implementación de la IA, así como las aplicaciones y limitaciones. Posteriormente, se entrevistó a expertos en recursos humanos y consultores de IA que ya habían adquirido experiencia de primera mano con los procesos de recursos humanos en un entorno de IA para descubrir la verdad de la aplicación de la IA dominante en el proceso de RRHH. Los principales hallazgos de esta tesis incluyen la derivación de una definición completa de IA en RRHH, así como el estado de las estrategias de adopción de aplicaciones de IA en RRHH. Como resultado adicional, se explora la utilidad y las limitaciones de los chatbots en el proceso de contratación en la India. Además, factores clave para transferir el conocimiento del proceso de implementación de IA a los gerentes y empleados de recursos humanos. Finalmente, se concluye identificando desafíos asociados con la implementación de IA en el proceso de recursos humanos y el impacto de COVID-19 en la implementación de IA. / [CA] El món és testimoni de nous avanços tecnològics, que afecten significativament les organitzacions en diferents departaments. La intel·ligència artificial (IA) és un d'aquests avanços que s'anuncia àmpliament com una tecnologia revolucionària en la gestió de recursos humans (HRM). Professionals i acadèmics han discutit el brillant paper de la IA en HRM. No obstant això, encara és escàs l'anàlisi profund d'aquesta tecnologia en el procés de HRM. Per tant, l'objectiu principal d'aquesta tesi és investigar l'estat de la IA en HRM i derivar factors clau d'implementació concrets. Primer, construint un marc acadèmic per a la IA en HRM; segon, analitzar l'aplicació de IA més utilitzada en el procés de recursos humans; tercer, identificar les formes òptimes de transferir el coneixement dels processos d'implementació de IA. La metodologia utilitzada per a la investigació es combina entre una revisió sistemàtica de la literatura i una tècnica d'investigació qualitativa. Com a base i mesura preparatòria per a abordar les preguntes d'investigació, es va dur a terme una extensa anàlisi de la literatura en el camp IA-HRM, amb un enfocament particular en les publicacions d'algorismes de IA en HRM, anàlisis de HR-Big data, aplicacions/soluciones de IA en HRM i implementació de IA. En la mateixa línia, l'autor va publicar articles en diverses conferències que van procedir a millorar la maduresa de les preguntes d'investigació. Amb base en aquest coneixement, els estudis publicats van illustrar la bretxa entre la promesa i la realitat de la IA en HRM, tenint en compte els requisits tècnics de la implementació de la IA, així com les aplicacions i limitacions. Posteriorment, es va entrevistar experts en recursos humans i consultors de IA que ja havien adquirit experiència de primera mà amb els processos de recursos humans en un entorn de IA per a descobrir la veritat de l'aplicació de la IA dominant en el procés de recursos humans. Les principals troballes d'aquesta tesi són la derivació d'una definició completa de IA en HRM, així com l'estat de les estratègies d'adopció d'aplicacions de IA en HRM. Com a resultat addicional, explore la utilitat i les limitacions dels chatbots en el procés de contractació a l'Índia. A més, factors clau per a transferir el coneixement del procés d'implementació de IA als gerents i empleats de recursos humans. També es van concloure els desafiaments associats amb la implementació de IA en el procés de recursos humans i l'impacte de COVID-19 en la implementació de IA. / [EN] The world is witnessing new technological advancements, which significantly impacts organizations across different departments. Artificial intelligence (AI) is one of these advancements that is widely heralded as a revolutionary technology in Human Resource Management (HRM). Professionals and scholars have discussed the bright role of AI in HRM. However, deep analysis of this technology in the HR process is still scarce. Therefore, the main goal of this thesis is to investigate the status of AI in HRM and derive concrete implementation key factors. Through, first, building an academic framework for AI in HRM; second, analyzing the most commonly used AI applications in HR process; third, identifying the optimal ways to transfer the knowledge of AI implementation processes. The methodology used for the investigation combines a systematic literature review and a qualitative research technique. As a basis and preparatory measure to address the research questions, an extensive literature analysis in the AI-HRM field was carried out, with a particular focus on publications of AI in HRM, HR-Big data analysis, AI applications/solutions in HRM and AI implementation. Along similar lines, the author published papers in several conference proceedings to improve the maturity of research questions. Based on this work, the published studies illustrate the gap between the promise and reality of AI in HRM, taking into account the requirements of AI implementation as well as the applications and limitations. Subsequently, HR experts and AI consultants, who had already gained first-hand experience with HR processes in an AI environment, were interviewed to find out the truth of the dominant AI's application in HR process. The main findings of this thesis are the derivation of a complete definition of AI in HRM as well as the status of the adoption strategies of AI applications in HRM. As a further result, it explores the usefulness and limitations of chatbots in the recruitment processes in India. In addition, derived the key factors to transfer the knowledge of AI implementation process to HR managers and employees. Challenges associated with AI implementation in the HR process and the impact of COVID-19 on AI implementation were also concluded. / Tuffaha, M. (2022). Adoption Factors of Artificial intelligence in Human Resource Management [Tesis doctoral]. Universitat Politècnica de València. https://doi.org/10.4995/Thesis/10251/185909 / TESIS / Compendio

The topic of this thesis is AI recruitment chatbots, digital discrimination, and data feminism (D´Ignazio and F.Klein 2020), where I aim to critically analyze issues of bias in these types of human-machine interaction technologies. Coming from a professional background of theatre, performance art, and drama, I am curious to analyze how using AI and social robots as hiring tools entails a new type of “stage” (actor’s space), with a special emphasis on social acting. Humans are now required to adjust their performance and facial expressions in the search for, and approval of, a new job. I will use my “theatrical glasses” with an intersectional lens, and through a methodology of cultural analysis, reflect on various examples of conversational AI used in recruitment processes. The silver bullet syndrome is a term that points to a tendency to believe in a miraculous new technological tool that will “magically” solve human-related problems in a company or an organization. The captivating marketing message of the Swedish recruitment conversational AI tool – Tengai Unbiased – is the promise of a scientifically proven objective hiring tool, to solve the diversity problem for company management. But is it really free from bias? According to Karen Barad, agency is not an attribute, but the ongoing reconfiguration of the world influenced by what she terms intra-actions, a mutual constitution of entanglement between human and non-human agencies (2003:818). However, tech developers often disregard their entanglement of human-to-machine interferences which unfortunately generates unconscious bias. The thesis raises ethical questions of how algorithmic measurement of social competence risks holding unconscious biases, benefiting those already privileged or those acting within a normative spectrum.