Maria Lucia Miotto Master Thesis Abstract in English Although more and more works are showing convergence between the many documents regarding the ethics of artificial intelligence, none of them has tried to explain the reasons for this convergence. The thesis here proposed is that the diffusion of these principles is due to the underlying action of an epistemic community that has promoted the spread and the adoption of these values. Then, through network analysis, this thesis describes the AI ethics epistemic community and its methods of value diffusion, testing for the most effective. Then, to test the first result, two case studies, representative of political opposites, the United States and the People Republic of China have been analysed to see which method of diffusion has worked the most. What seems evident is that scientific conferences remain a primary factor in the transmission of knowledge. However, particular attention must also be given to the role played by universities and research labs (also those of big tech-companies) because they have revealed to be great aggregators for the epistemic community and are increasing their centrality in the network.

Yes / Since the 1980s, the digital revolution has been both a negative and positive force. Within a few weeks of the Covid-19 outbreak, lockdown accelerated the adoption of digital solutions at an unprecedented pace, creating unforeseen opportunities for scaling up alternative approaches to social and economic life. But it also brought digital risks and threats that placed new demands on policymakers. This article assembles evidence from different areas of social science expertise about the impacts of Covid-19 in digitised societies and policy responses. The authors show how the pandemic supported changes in data collection techniques and dissemination practices for official statistics, and how seemingly insuperable obstacles to the implementation of e-health treatments were largely overcome. They demonstrate how the ethics of artificial intelligence became a primary concern for government legislation at national and international levels, and how the features enabling smart cities to act as drivers of productivity did not necessarily give them an advantage during the pandemic. At the micro-level, families are shown to have become ‘digital by default’, as children were exposed to online risks and opportunities. Globally, the spread of the pandemic provided a fertile ground for cybercrime, while digital disinformation and influencing risked becoming normalised and domesticated.

Medellivslängden i Sverige har, under de senaste decennierna, ökat. Det leder till att sjukvårdsköerna och belastningen för vårdpersonalen förändras. Det i sin tur innebär att fortsatt utveckling av hälso- och sjukvård måste prioriteras. En viktig faktor för att kunna utveckla arbetssätten inom sjukvården är digitalisering.  Syftet med studien är att undersöka vilka utmaningar som kan uppstå i samband med användning av AI- hjälpmedel som beslutsstöd vid diagnostiserings-processen i sjukvården.  Studien startade med litteraturgenomgång av tidigare forskning. Studien genomfördes med en kvalitativ ansats där datainsamlingsmetoden är semistrukturerade individuella intervjuer. Analysen genomfördes med strukturen tematisk analysmetod. Studiens fem teman är: Arbetssätt; kompetensutveckling; partiska algoritmer; ansvarstagande och överdiagnostisering.  Studiens resultat indikerar att det finns olika utmaningar vid användning av AI inom vården. I den här studien har följande utmaningar påvisats: arbetssätt- exempelvis nya rutiner, kompetensutveckling- exempelvis mindre tid för utvecklingsarbete, ansvarstagande - till exempel vem som ansvarar för resultatet och överdiagnostisering – vilket till exempel kan vara multipla diagnoser som kan vara ofarliga för patienten. / The average life expectancy in Sweden has, in recent decades, increased. This leads to changes in the healthcare queues and the burden on the healthcare staff. This in turn means that continued development of healthcare must be prioritized. An important factor in being able to develop working methods in healthcare is digitization. The purpose of the study is to investigate which challenges may arise in connection with the use of AI- powered decision making in the diagnosis process in healthcare. The study started with a literature review of previous research. The study was carried out with a qualitative approach where the data collection method is semi-structured individual interviews. The analysis was carried out using the structured thematic analysis method. The study's five themes are: Work methods; skills development; biased algorithms; taking responsibility and overdiagnosing. The study's results indicate that there are various challenges when using AI in healthcare. In this study, the following challenges have been demonstrated: working methods - for example new routines, competence development - for example less time for development work, taking responsibility - for example who is responsible for the result and overdiagnosis - which can for example be multiple diagnoses that can be harmless to the patient.

Many, if not all people have online social accounts (OSAs) on an online community (OC) such as Facebook (Meta), Twitter (X), Instagram (Meta), Mastodon, Nostr. OCs enable quick and easy interaction with friends, family, and even online communities to share information about. There is also a dark side to Ocs, where users with malicious intent join OC platforms with the purpose of criminal activities such as spreading fake news/information, cyberbullying, propaganda, phishing, stealing, and unjust enrichment. These criminal activities are especially concerning when harming minors. Detection and mitigation are needed to protect and help OCs and stop these criminals from harming others. Many solutions exist; however, they are typically focused on a single category of malicious intent detection rather than an all-encompassing solution. To answer this challenge, we propose the first steps of a framework for analyzing and identifying malicious intent in OCs that we refer to as malicious mntent detection framework (MIDF). MIDF is an extensible proof-of-concept that uses machine learning techniques to enable detection and mitigation. The framework will first be used to detect malicious users using solely relationships and then can be leveraged to create a suite of malicious intent vector detection models, including phishing, propaganda, scams, cyberbullying, racism, spam, and bots for open-source online social networks, such as Mastodon, and Nostr.