IFP-nytt 26. februar

• Kj?rven, Marte Eidsand; Gj?steen, Kristian & W?rstad, Tone Linn (2026). Safe and inclusive or unsafe and discriminatory? European digital identity wallets and the challenges of ¡®sole control¡¯. Computer Law & Security Review. ISSN 2212-473X. 60, s. 106235¨C106235. doi: 10.1016/j.clsr.2025.106235. Fulltekst i vitenarkiv

Se alle arbeider i NVA

• R?s?g, Erik (2052). Pantsettelse og kreditorbeslag av klimakvoter. Fulltekst i vitenarkiv
• Irgens-Jensen, Harald (2026). EU-domstolen presiserer forbrukerbegrepet ¨C C-197/24. Nytt i privatretten : nyhetsbrev for informasjon p? det privatret. ISSN 1501-9594. 2. Fulltekst i vitenarkiv Vis sammendrag

  Ny avgj?relse fra EU-domstolen om hvem som kan regnes som "forbruker"

• Salvesen, Sverre Magnus Bergslid (2026). Studenter vurderer masses?ksm?l mot flyskolens styre ¨C NRK Vestfold og Telemark ¨C Lokale nyheter, TV og radio. [Internet]. NRK. doi: https:/www.nrk.no/vestfoldogtelemark/studenter-vurderer-massesoksmal-mot-flyskolens-styre-1.17746461. Fulltekst i vitenarkiv Vis sammendrag

  Inkassokravene rant inn, men studentene ble likevel bedt om ? betale for et nytt semester. N? vurderer studenter og fagforeningen Parat om de kan stille styret personlig ansvarlig for milliontapene.

• Stenseth, Geir (2026). Store endringer: - Helt annen aggressivitet. [Journal]. Dagbladet. Fulltekst i vitenarkiv Vis sammendrag

  Innvandringspolitiet ICE anklages for ? operere som en milits under Trump-administrasjonen. Hva har de egentlig lov til ? gj?re?

• Salvesen, Sverre Magnus Bergslid (2026). Pilot flight academy gikk konkurs: Chris vasker flyplassen om natta for ? fullf?re pilotstudiene ¨C NRK Vestfold og Telemark ¨C Lokale nyheter, TV og radio. [Internet]. NRK. doi: https:/www.nrk.no/vestfoldogtelemark/pilot-flight-academy-gikk-konkurs_-chris-vasker-flyplassen-om-natta-for-a-fullfore-pilotstudiene-1.17725961. Fulltekst i vitenarkiv Vis sammendrag

  Studentene trodde pengene de betalte p? forskudd til Pilot Flight Academy havnet p? sperret konto. Det skjedde aldri. ¨C Kan v?re kontraktsbrudd, mener jusekspert.

• Stenseth, Geir (2026). Slik kan Trump ta Gr?nland. [Journal]. Aftenposten. Fulltekst i vitenarkiv Vis sammendrag

  Det har fremst?tt som et utenkelig scenario. N? snakkes det h?yt om. Men hvordan kan Trump ta Gr?nland?

• Stenseth, Geir (2026). Kan bli tollsmell for Trump i H?yesterett. [Journal]. E24. Fulltekst i vitenarkiv
• Stenseth, Geir (2026). Dette kan stanse ham. [Journal]. Dagbladet. Fulltekst i vitenarkiv
• Palmieri, Alice (2025). Systemic Risk Reporting under the DSA: A Study of Seven VLOPs in their Operationalisation of the DSA Systemic Risk Management. Universitetet i Oslo. Fulltekst i vitenarkiv
• Fan, Yujie (2025). ¡°Manipulative by Design¡±: Regulating AI-powered Social Media Recommender Systems in the EU. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  This thesis investigates the manipulative nature of social media recommender systems (SMRS) and examines the extent to which the latest EU regulation addresses the harm of manipulative SMRS on individuals and democracies. This includes the ethical challenges that SMRS pose to individual autonomy and democratic processes, how the most recent applicable regulations in the EU (Digital Service Act and AI Act) address these ethical challenges and risks, and their gaps and limitations.

• Mctavish, Charlotte Naomi (2025). Smartphone and Social Media Screening Legislation and Practices in the Asylum Procedure: Violating the right to privacy or a necessary measure to establish identity and safeguard national security? Universitetet i Oslo. Fulltekst i vitenarkiv
• Nicosia, Lucrezia (2025). Prospects for a Coordinated Global Approach in the AI Landscape: A Comparative Analysis of the EU, US, China, and UAE. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  The rapid growth of artificial intelligence (AI) has produced a fragmented global regulatory landscape, highlighting the need for coordinated governance. This thesis addresses the prospects for achieving a globally coordinated approach to AI governance by comparing the frameworks of the European Union (EU), United States (US), China, and United Arab Emirates (UAE). It employs a comparative legal methodology using the common-core method and qualitative analysis of each jurisdiction¡¯s AI laws, national strategies, and policy instruments. The findings reveal divergent regulatory models shaped by distinct legal cultures and priorities: the EU¡¯s comprehensive, rights-driven and risk-based regulations, the US¡¯s innovation-centric and decentralized approach, China¡¯s state-driven and techno-nationalist governance, and the UAE¡¯s strategic vision of becoming a global hub for AI development and deployment. Despite some convergence through international guidelines and standards (soft law), the research observes that geopolitical and economic forces ¨C including AI nationalism and strategic competition ¨C complicate efforts toward formal harmonization. International law initiatives remain nascent and limited, while non-binding frameworks provide only partial alignment. Conclusions indicate that full global regulatory uniformity remains unlikely in the near term. However, incremental steps toward regulatory interoperability are feasible, as jurisdictions collaborate on shared principles and best practices. This study underscores both the urgent impetus for global AI governance and the significant limitations to achieving a truly unified framework, offering a foundation for future dialogue on global AI regulation.

• Gutierrez, Jamila Shadjealom Mendoza (2025). Information and Communication Technology Law. Universitetet i Oslo. Fulltekst i vitenarkiv
• Liutyi, Sviatoslav (2025). Artificial Intelligence-Generated images under European Union Copyright law. Universitetet i Oslo. Fulltekst i vitenarkiv
• Langen, Darleen (2025). Early Warning Systems and automated decision making in peacebuilding processes - Regulatory challenges and a case study of the Violence & Impacts Early-Warning System (ViEWS) under the EU AI Act. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  As a contemporary technology, artificial intelligence (AI) has already permeated numerous sec-tors, consequently giving rise to a corpus of legal regulations about its deployment in specific domains. These include the utilisation of AI in contexts such as warfare, national security, healthcare, and surveillance. In particular, the first-mentioned areas of application regard AI as a tool that can be harmful. In contrast to the extensive research and development that has been conducted on the use of AI for these applications, there has been relatively little attention di-rected towards exploring its potential as a tool for building peace. Meanwhile, an interdiscipli-nary field of research has emerged, designated as PeaceTech, which is concerned with the utili-sation of technology to foster peace. Peacebuilders traditionally analysed new technologies in terms of their effectiveness for peacebuilding. ¡°PeaceTech is an umbrella term for technologies (software and hardware) that are be-ing developed and used in efforts to prevent or end cycles of violence in society, building and sustaining peace.¡± This also includes the use of AI for peace. There are already some models that are used for this purpose, the so-called Early Warning Systems (EWS). EWSs collect data, assess a risk, provide information, and recommend actions to relevant stakeholders. They are used by international organisations, research institutions, and Civil Society Organisations (CSOs) as well as bilateral agencies. Worth mentioning here are the Violence & Impacts Early-Warning System (ViEWS) operated by the Peace Research Institute Oslo (PRIO), which will be the subject of a case study in this thesis. Recent developments in machine learning signify a paradigm shift in the realm of strategic warning systems. This technological advancement holds immense potential for enhancing the efficiency and precision of such systems, thereby aiming for optimal decision-making process-es in complex and dynamic environments. The potential benefits of such improvements include the provision of real-time alerts to analysts and policymakers regarding political events in areas that have not been previously monitored, as well as the faster identification of military aggres-sion in regions already subject to monitoring. The primary advantage of these enhancements is that they enable policymakers to have more time to act. This allows for the reallocation of re-sources and a more informed consideration of the consequences of various risk-mitigation measures. Such measures may include those that form part of a deterrence strategy or that miti-gate the impact of a pending attack. Some voices say that the integration of AI has the potential to close the warning-response gap. Notwithstanding its auspicious potential, AI poses a fundamental dilemma for peacebuilders and human rights defenders working in PeaceTech. To leverage technology for positive change, it is equally crucial to confront the risks it poses. While AI can support peace, justice, and equality, it also harbours the potential to undermine human rights, deepen inequalities, and destabilise peace processes. It is therefore vital for practitioners operating within PeaceTech to navigate this tension in order to ensure that technological innovation aligns with ethical and humanitarian principles. These developments are prompting legislators around the world to enact regulations governing the utilisation of AI. A considerable number of regulations have already been enacted by various legislative bodies, including states, supranational organizations, and private entities. One of the recently introduced legislative measures is the AI Act (AIA), which was enacted by the Europe-an Union (EU) and argues for a human-centric approach for future developments in AI. The AIA has provisions that have an anthropocentric orientation. In particular, Article (Art.) 14 and recital (Rec.) 27 are references within this law that explain the human-centred approach in more detail. AI technology should serve people, respect human dignity and personal autonomy, and must be appropriately controlled and monitored by humans. Human supervision is mandatory for high-risk systems.

• Pearson-Woodd, Anna Renata (2025). Automated individual decision-making by AI systems: A critical analysis of Article 22 of the EU GDPR and Article 86 of the EU AI Act. Universitetet i Oslo. Fulltekst i vitenarkiv
• Miltsov, Maja Ottersen (2025). Will the AI Act make Chatbots Trustworthy Again? Universitetet i Oslo. Fulltekst i vitenarkiv
• Sidiropoulou, Alexandra (2025). AI Compliance and Sustainability: The EU AI Act¡¯s Role and Outlook. Universitetet i Oslo. Fulltekst i vitenarkiv
• Halkjelsvik, Peter Eik?s (2025). Tilpassing av skytenester til offentlege verksemder sine behov - Ein casestudie av Statnett si tilpassing av skytenesten Omega 365. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  Denne oppg?va omhandlar Statnett si tilpassing av skytenesten Omega 365. Oppg?va granskar korleis Statnett har tilpassa skytenesten til sine behov, kva utfordringar dei har st?tt p? i denne prosessen og om desse utfordringane stemmer overens med forskinga si skildring av utfordringar i anskaffing og tilpassing av skytenester. Meir spesifikt fokusere oppg?va p? kva innverknad etterlevese av strenge krav til sikkerheit og personvern kan f? p? tilpassinga av skytenester i offentleg forvalting.

• Mellingen, Aksel (2025). Personvern og integrert kunstig intelligens: En studie av arbeid med etterlevelse av personvernlovverket ved implementering av Microsoft 365 Copilot. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  Det integrerte kunstig intelligens-systemet Microsoft 365 Copilot kan gi betydelige gevinster for virksomhetene som tar det i bruk, og en rekke norske virksomheter vurderer eller har allerede implementert det for sine ansatte. Det gjenst?r likevel ? avklare om m?ten Copilot behandler personopplysninger p? ivaretar personvernet til de registrerte. I denne oppgaven kartlegger jeg med en dokument- og litteraturstudie hvordan personopplysninger vil bli behandlet i Copilot. Gjennom en forenklet rettsdogmatisk metode vurderer jeg s? hvordan disse behandlingene kan ses i lys av personvernprinsippene og tilknyttede lovbestemmelser fra det sentrale personvernlovverket. Jeg gj?r ogs? en casestudie av Arbeids- og velferdsetaten (Nav) sin implementering av Copilot, der jeg unders?ker hvordan virksomheter kan g? frem for ? ta i bruk Copilot p? en m?te som sikrer etterlevelse av personvernlovverket. I oppgaven kommer jeg frem til at behandlingsopplegget Microsoft legger opp til i Copilot kan medf?re utfordringer mot flere av personvernprinsippene, og da s?rlig prinsippet om lovlighet, rettferdighet og ?penhet, form?lsbegrensningsprinsippet, dataminimeringsprinsippet og riktighetsprinsippet. Ivaretagelsen av andre prinsipper, som prinsippet om konfidensialitet og integritet, stiller ogs? h?ye krav til forvaltningen av KI-systemet. Nav m?tte flere av de samme personvernutfordringene i sitt arbeid med Copilot. Deres erfaringer tilsier at god kontroll p? virksomhetens datagrunnlag, en stegvis implementeringsprosess, grundige testing av Copilot i egne systemer og god oppl?ring av ansatte er viktige tiltak for ? sikre personvernet i bruken av Copilot. Tverrfaglig 澳门葡京手机版app下载 internt i etaten og erfaringsutveksling mellom forskjellige offentlige virksomheter ble ogs? l?ftet frem som nyttige tiltak i Navs arbeid med Copilot.

• Vasilev, Petr (2025). Universell utformede IKT-l?sninger ¨C et svar p? digital utenforskap? Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  Denne masteroppgaven unders?ker sammenhengen mellom digitalt utenforskap og universell utforming av IKT-l?sninger, med s?rlig fokus p? regulering, h?ndheving og praktisk implementering i Norge. Digitalisering av offentlig sektor gir bedre tilgang og effektivitet i tjenesteyting, men skaper samtidig risiko for at enkelte brukergrupper ekskluderes, noe som resulterer i digitalt utenforskap. Oppgaven identifiserer sentrale barrierer som tilgang til digitale verkt?y, digitale ferdigheter, byr?kratisk kompetanse, helseproblemer og spr?klige utfordringer og ser hvordan universell utforming av IKT-l?sninger brukes for ? senke barrierene. Ved bruk av litteratur- og dokumentstudier, juridisk metode, og et semistrukturert intervju med en ansatt i tilsynet for universell utforming (uu-tilsynet), analyserer oppgaven hvordan dagens reguleringsrammer og praksiser bidrar til ? redusere digitalt utenforskap. Spesielt ser oppgaven n?rmere p? et konkret tilsyn utf?rt hos Utdanningsdirektoratet, som illustrerer hvordan risikobasert tilsyn anvendes i praksis. Oppgavens funn viser at dagens reguleringer og standarder legger et viktig grunnlag og f?lger beste praksis, men det gjenst?r likevel utfordringer knyttet til koordinering mellom ulike offentlige akt?rer, ulikheter i regelverk mellom offentlig og privat sektor, og begrenset kapasitet hos tilsynsmyndighetene. Analysen peker p? behovet for styrket 澳门葡京手机版app下载 mellom kultur- og digitaliseringssektorer.

• Andersen, Mathias Kristian Emil (2025). Innkj?psanalyseverkt?y i kommunal sektor. Universitetet i Oslo. Fulltekst i vitenarkiv
• Mikelyte, Patricija (2025). Demonstrating compliance with Binding Corporate Rules: Blockchain technology as an accountability tool. Universitetet i Oslo. Fulltekst i vitenarkiv
• Yoseph, Esayas, Samson & Axel, Rantanen, (2025). Compliance Through Technology and Innovation: Evaluating PETs for Implementing Technical Measures in Big Data Analytics in Accordance with Article 25 GDPR. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  -

• Weinzierl, Prof. Quirin; Tosoni, Prof. Luca & Abuchi, Obiekwe Nuel (2025). Securing the Grid: Digital Resilience and Liability in EU Smart Grid and SCADA Supply Chains. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  Summary This thesis examines how the European Union¡¯s cybersecurity framework addresses supply-chain security and digital resilience in IoT-enabled smart grids and SCADA systems within the renewable energy sector. It focuses on the interaction between the NIS2 Directive,CRA and CER, together with the Product Liability Directive and GDPR, analysing how these instruments collectively redistribute legal liability and operational accountability across the energy value chain. The study adopts a doctrinal legal methodology, analysing EU legislation, policy documents, standards, and secondary literature. Rather than assessing the instruments in isolation, the thesis takes an integrated approach, examining their cumulative effects on risk management, lifecycle security obligations, and governance structures. It finds that the EU has moved from a voluntary, operator-centric cybersecurity model toward a legally enforceable regime of shared responsibility. The CRA shifts accountability upstream to manufacturers and other economic operators, while NIS2 embeds cybersecurity within corporate governance through executive-level obligations. CER further situates digital resilience within a broader framework for protecting critical infrastructure. The thesis also analyses how these obligations are likely to be transposed and applied in Norway under the EEA Agreement. It identifies key legal and institutional challenges, including fragmented supervisory competences, tensions between EU information-sharing requirements and national security protections under the Security Act, and the risk of overlapping or inconsistent reporting obligations. The study argues that effective transposition will require coordinated governance and functional equivalence rather than formal legislative alignment alone. While regulatory reform addresses many of these issues, the thesis concludes that achieving durable resilience will depend on integrating legal obligations with operational practices, technological support mechanisms, and a shared understanding of resilience as both a market and constitutional value.

• Berg, ?rstavik Inger & Hrithik, Verma (2025). Governing the grid - Data Control in Decentralized Energy Systems. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  As the energy sector transitions from centralized utilities to decentralized Peer to Peer systems, the legal status of the resulting high resolution data flows remains ambiguous. This thesis investigates the research question: ¡°Who owns and controls energy data in decentralized systems, and under what legal frameworks is this control allocated?¡± Adopting a doctrinal and comparative methodology, the study analyzes the intersection of the GDPR, the EU Data Act, and the Norwegian energy law. It challenges the traditional property based view of data, arguing that the concept of ¡°data ownership¡± is legally flawed and practically insufficient to govern the complex, multi actor environment of the PriTEM model. Instead, the thesis identifies a ¡°role-based model of distributed control¡± where rights are functionally allocated, prosumers hold decisional control, DSOs hold custodial control, and platforms hold functional control. Through a detailed legal analysis of the PriTEM architecture, the study demonstrates that compliance is achieved not by centralizing ownership, but by orchestrating these roles through Joint Controllership agreements (Article 26 GDPR) and Privacy by Design. The thesis concludes that while the current legal framework is robust, it creates significant administrative friction for energy communities, necessitating standardized liability models and automated governance, tools to ensure the law enables , rather than inhibits, the green transition. Oslo, December 2025

• L¨²cia., Raposo, Vera & Iris., Tuijaerts, (2025). GDPR: Barrier or Bridge? The right of access to donor identification after abolition of anonymity concerning gamete and embryo donation in Belgium. Universitetet i Oslo. Fulltekst i vitenarkiv
• Tobias, Mahler, & Thi, Tran, Giang Thuy (2025). Quality in AI-Based Translation: Operationalizing the Language Act¡¯s Requirement of Parallel Language Use. Universitetet i Oslo. Fulltekst i vitenarkiv Vis sammendrag

  This thesis examines how AI-based translation tools can be assessed for use in the public sector in light of the Norwegian Language Act¡¯s requirement of parallel language use. Although Bokm?l and Nynorsk are legally equal written standards, supervisory reports show that Nynorsk users still face clear disadvantages in digital communication. At the same time, public bodies increasingly rely on AI-assisted tools to meet expectations of efficiency. This creates a tension between technological convenience and the democratic aim of securing equal access to information. The thesis addresses this gap by developing a framework for evaluating the trustworthiness of AI-generated translations between Bokm?l and Nynorsk. In this context, trustworthiness is understood through practical dimensions of linguistic quality, such as accuracy, consistency and the extent to which the two language versions function equally well for users. The analysis draws on insights from linguistics, translation studies and selected international frameworks. Since the Language Act provides no explicit guidance on translation quality, these supplementary perspectives are used to clarify how central aspects of quality and clarity can be assessed in a structured and predictable manner. Two criteria are proposed. The first concerns linguistic accuracy, consistency and functional equivalence. Whether the translated text conveys the intended meaning, uses terminology in a coherent way and enables the recipient to understand and act on the information. The second concerns understandability and readability. Focusing on how the structure, flow and explanation of concepts support practical comprehension in administrative contexts. Each criterion is operationalized through indicators and assessed at micro, meso and macro levels, allowing both sentence-level precision and broader structural considerations to be evaluated. The thesis argues that public bodies cannot rely on AI-generated Nynorsk without systematic quality assurance. Parallel language versions must function equally well in practice if they are to fulfil the purpose of the Language Act. Inconsistent or unclear translations risk weakening citizens¡¯ ability to navigate public information and may undermine the intention of linguistic equality. By offering a structured and context-sensitive framework, the thesis aims to support more predictable, transparent and responsible use of AI-based translation tools in public administration.

• Tobias, Mahler, & Marie, Heggland, (2025). Regulating Open-Source Software under the EU Cyber Resilience Act. Assessing whether the CRA strengthens cybersecurity and collaboration in the EU open-source ecosystem or unintentionally hinders open-source development. Universitetet i Oslo. Fulltekst i vitenarkiv

Se alle arbeider i NVA