Our work
Here you will find all our work and everything others write about us. We are constantly working to produce new content, including notes, consultation input, chronicles and debate posts.
Highlighted
Nyeste
Priority areas
All publications
Secure KI requires more than research
Norway needs a government body that can coordinate the work of AI security.
The aid works — when someone asks questions
Over the past 15 years, Noreg has donated funds to more than 100 American research institutions, think tanks and consulting firms, including the International Peace Institute (IPI), without anyone questioning its effectiveness. The problem, however, is the process, not that aid does not work.
Technology-driven Biological Threats
Rising risks driven by advances in biotechnology and artificial intelligence
How to Handle the Greatest Threats
Gyldendal's "Total Preparedness": In a short time, the security policy situation in Norway and Europe has intensified. While war and geopolitical upheavals once again dominate the news, preparedness challenges resulting from extreme weather, digital vulnerability, and complex threats have increased in both scope and complexity.
Someone needs to test the KI models Norway relies on
Norway is lagging behind in efforts to secure artificial intelligence. We propose a national security body.
Now the shittification comes to KI
ChatGPT now provides you with custom advertising when you ask for life advice. Welcome to the shittification of KI.
AI development, threats and Norwegian opportunities
Proposals for investments and institutional innovations
Helvetes AI-agenter
AI-agenter skjuler seg bak språk som «assistenter», «copiloter» og «verktøy». Den reelle effekten er å organisere arbeid bort fra mennesker. Verdiskapingen skjer nettopp fordi dette ikke er tydelig for dem som rammes.
Bruk gjerne KI til å skrive tekst, men det er kritisk viktig at det finnes en avsender
I offentligheten er det likegyldig om man får skrivehjelp av en venn, enten vennen er av silisium eller kjøtt og blod.
Data discovery — a tax incentive for sharing data
Norway can't speed up KI without better data, and no one has incentives to make data shareable. That's why we need a tax incentive for data quality and sharing.
Media coverage
Can artificial intelligence withstand a scorching hot planet?
Terje Erikstad writes about threats from artificial intelligence and highlights Langsikt's seminar on biological threats.
Norad risks undermining what they aim to strengthen
Digni responds to Cindy Robles' op-ed on Norad's proposed changes to civil society support.
– We must do less of everything that dilutes aid
Senior Advisor Cindy Robles was interviewed about how she would have reformed Norwegian aid if she had been in charge.
More than the number of mosquito nets and vaccines
Digni's CEO is interviewed about aid in Dagen, and Eirik Mofoss' criticism of the Ministry of Foreign Affairs is mentioned.
AI, blunders, and trust.
Expert Aksel Braanen Sterri interviewed on NRK Radio about an AI blunder during the national anthem.
Norwegian authorities should not legitimize effective altruism.
Jörn Klein criticizes Norwegian authorities for participating in a Langsikt seminar, citing Langsikt's informal role in the effective altruism movement. Eirik Mofoss and Aksel Sterri respond in their own article.
Short-term plan
Morten Jerven criticizes Eirik Mofoss' and Cindy Robles' article in NRK Ytring on cash transfers.
Experts concerned about the development: – The threat landscape is changing
Those working to protect Norway from biological threats are concerned. Artificial intelligence and new biotechnology may be changing the rules of the game. A report following the Langsikt seminar on biological threats in a new era.
The world's most powerful cyber weapon is here. His name is Claude.
Incoming AI advisor Anders Eidesvik writes about Claude as a superhacker in Altinget.
What happens at the intersection of advanced machine learning, applied ethics, and sociopolitical transformation?
Advisor and AI researcher Preben Ness visited Skravleklassen to discuss AI.
