Insights

Analysis on AI risk and resilience.

Perspectives on AI governance, agentic AI risk, cybersecurity and Europe’s resilience.

AI Risk2 November 20267 min read

Is Europe ready for what comes next? The founding story of NIARR

Why the Nordic Institute for AI Risk and Resilience was founded: an AI ecosystem working in silos, the interdisciplinary spirit of Dartmouth, the AI risks of 2026, and what Europe needs not only to compete in the AI era, but to defend itself.Read the article

AI Risk22 October 20266 min read
PHP2Shell Incident: Finding Bugs Is Getting Easier. Stopping Attacks Is the New Sport.
Two critical vulnerabilities in PHP put websites around the world at risk, years after the weaknesses first appeared. AI-assisted research showed how they could be turned into working attacks, and the time between discovering a weakness and protecting the systems that depend on it is the real challenge.Read the article
AI Governance4 August 20264 min read
The Hugging Face – OpenAI Cyber Incident: Agentic AI Risk has now entered a new phase
The Hugging Face–OpenAI cyber incident may prove to be one of the most important public precedents for understanding agentic AI risk. This article examines what happened, why it matters from legal, cybersecurity, and AI governance perspectives, and why boards, investors, and policymakers should begin treating AI governance as a strategic capability rather than a compliance exercise.Read the article
Article18 June 20268 min read
Frontier AI Risk Has Moved Beyond National Security to a Critical Strategic Dependency for the EU
This analysis explores why AI risk has moved from a national security risk to a question of European sovereignty. On 12 June 2026, the U.S. administration restricted foreign access to Anthropic’s most advanced AI models, Fable 5 and Mythos 5, including access by foreign nationals inside the United States and foreign national Anthropic employees. Access to frontier AI has become a geopolitical decision taken outside Europe, with direct consequences for European industries. Frontier AI has become critical infrastructure. This is no longer only about safe AI. It is also about market access and gatekeeping, technological leadership, scientific innovation, international trade, market resilience, information integrity, and Europe’s long term ability to compete globally. It also reflects why I have decided to become more active in the political arena at both the Swedish and European level. Europe cannot afford to be late on this one.Read the article
AI Risk13 May 20266 min read
Anthropic’s “Mythos” Marks the Rise of Advanced Offensive AI Capabilities as AI Sweden Warns of Urgent National Security Exposure
AI Risk Has Entered a New Critical Phase. This is no longer only a cybersecurity discussion. It increasingly intersects with operational resilience, supplier dependencies, governance maturity, and critical infrastructure preparedness and strongly challenges the resilience of the current outdated and manual governance structures. This article explains the urgency and clarifies what is changing for Sweden, the Nordics, and Europe.Read the article
AI Governance8 February 20266 min read
Shadow AI: A Critical Organisational and National Security Risk for 2026
This article examines the rapid normalisation of AI tools that secretly record meetings, now widely marketed on platforms such as Instagram, TikTok, and Facebook. It outlines governance risks, organisational responsibilities, regulatory expectations, and why shadow AI is a present, not future, challenge.Read the article
Cybersecurity15 January 20265 min read
Sweden’s Cybersecurity Act has now entered into force
From 15 January 2026, Sweden’s Cybersäkerhetslagen implements NIS2 and reshapes how organisations must govern cybersecurity and digital risk. This article explains who is in scope, key governance expectations, and why structured risk management and accountability are now essential.Read the article
Compliance19 August 20252 min read
ISO 42001: How to Gain Competitive Advantage by Leveraging AI Governance
ISO 42001 is the world’s first standard for AI management systems, designed to turn AI governance into a competitive advantage. Building on ISO 27001, it helps organisations move from principles to practice, strengthening resilience, demonstrating regulatory readiness, and proving that AI risks are actively managed.Read the article
AI Governance23 July 20256 min read
The EU GPAI Code of Practice: Advancing Trust Through Accountability
The EU’s new GPAI Code of Practice is ready for adoption, just ahead of the 2 August 2025 enforcement date. For organisations building or integrating GPAI models, it’s an opportunity to lead with confidence, gaining a one-year enforcement buffer, clear operational guidance, and a visible mark of accountability that builds trust with regulators, clients, and partners.Read the article
AI Governance10 June 20253 min read
Japan Passes Landmark AI Law: A New Governance Model for the Age of Artificial Intelligence
On May 28, 2025, Japan enacted its first national AI legislation, the AI Promotion Act; introducing a pioneering governance model that prioritizes structured collaboration, public transparency, and adaptive policy-making over traditional enforcement mechanisms. Find out more about the Japanese approach to AI governance in this article, and what makes it different.Read the article
ISO3 June 20253 min read
The New ISO standard for AI Impact Assessment: ISO/IEC 42005 Published in May 2025
Released in May 2025, ISO/IEC 42005 is the newest European AI standard focused on real-world fundamental rights impact assessment. Paired with ISO 42001, it marks a major shift in global AI governance.Read the article
AI Governance21 May 20253 min read
The Take It Down Act: U.S. Federal Law Addressing Non-Consensual Intimate Imagery and AI-Generated Deepfakes
Enacted on May 19, 2025, the Take It Down Act criminalizes non-consensual intimate images, including AI-generated deepfakes, and requires platforms to remove such content within 48 hours. It introduces “digital forgery” as a legal term and imposes strict compliance and penalties.Read the article
EU AI Act12 May 20253 min read
How to Operationalize Human Oversight in HR: Key Steps for AI Governance and Privacy
Meaningful human oversight is a legal obligation under the EU AI Act for high-risk AI systems, requiring active, informed intervention by individuals with the authority and in-depth understanding of AI systems (AI literacy) to override decisions and identify compliance risks. This article explores how oversight must function in practice and how it intersects with automated decision-making under GDPR Article 22.Read the article
Article8 May 20254 min read
Shadows of Surveillance: How AI Exploitation Undermines National Security and Human Rights
Artificial intelligence is rapidly reshaping the world, but without robust AI governance and data protection laws, the consequences can be devastating for democratic societies. From AI-driven phishing attacks targeting seniors to AI-generated propaganda manipulating public opinion, the risks to national security are escalating. This article explores how AI exploitation is endangering data sovereignty, spreading disinformation, and deepening ideological divides, and what governments must do to protect vulnerable populations and maintain public trust.Read the article
Artificial Intelligence8 September 20243 min read
The EU AI Act: World’s first comprehensive AI Law as of August 2024
Gain insight into the AI Act’s objectives and requirements, and learn how to navigate AI regulation, ensuring compliance and mitigating potential legal risks for businesses.Read the article

Coming to Insights

Coming soon
Resources
Guides and tools on AI risk and resilience, gathered in one place.
Coming soon
White papers
In-depth papers from NIARR and its members.
Coming soon
Laws and regulation
The rules that shape AI, made easy to find and follow.
Coming soon
Related news
Selected news on AI risk and resilience from other sources.

Research together.

We welcome conversations with researchers, institutions and practitioners who share these questions.

Propose a collaboration

Analysis on AI risk and resilience.

Is Europe ready for what comes next? The founding story of NIARR

PHP2Shell Incident: Finding Bugs Is Getting Easier. Stopping Attacks Is the New Sport.

The Hugging Face – OpenAI Cyber Incident: Agentic AI Risk has now entered a new phase

Frontier AI Risk Has Moved Beyond National Security to a Critical Strategic Dependency for the EU

Anthropic’s “Mythos” Marks the Rise of Advanced Offensive AI Capabilities as AI Sweden Warns of Urgent National Security Exposure

Shadow AI: A Critical Organisational and National Security Risk for 2026

Sweden’s Cybersecurity Act has now entered into force

ISO 42001: How to Gain Competitive Advantage by Leveraging AI Governance

The EU GPAI Code of Practice: Advancing Trust Through Accountability

Japan Passes Landmark AI Law: A New Governance Model for the Age of Artificial Intelligence

The New ISO standard for AI Impact Assessment: ISO/IEC 42005 Published in May 2025

The Take It Down Act: U.S. Federal Law Addressing Non-Consensual Intimate Imagery and AI-Generated Deepfakes

How to Operationalize Human Oversight in HR: Key Steps for AI Governance and Privacy

Shadows of Surveillance: How AI Exploitation Undermines National Security and Human Rights

The EU AI Act: World’s first comprehensive AI Law as of August 2024

Coming to Insights

Resources

White papers

Laws and regulation

Related news

Research together.