Abstract
The risk-management toolkit most operators still use was built for a world of stochastic failure, but today the shift to deliberate, state-backed adversaries breaks the core assumptions. Probability is no longer a property of the world derived from history, but a property of someone else’s intent; redundancy that halves the risk of random failure offers no protection when an adversary hunts for the shared weakness, and “low probability × high impact = accept” decisions made years ago are now wrong but rarely revisited. The talk layers on three accelerants: the move from digital complements to digital dependencies, AI that compresses the attack timeline so that response plans built on business-day clocks become irrelevant, and AI agents that act as insiders whose attack surface is defined by permissions rather than model capability. A live dependency measurement of europa.eu and regeringen.se illustrates that “sovereignty” is not captured by origin requirements but by dependency analysis. No organisation can secure itself in isolation, since everyone is one link in chains that run through others.
Recording
Speaker
Fredrik Lindeberg
Fredrik is currently working as Chief Information Security Officer at Netnod. His role is focused on internal processes, public policy development and the implications of policy on Netnod’s services. In 2021, Fredrik defended his PhD-thesis, Coordinating the Internet: Thought styles, technology and coordination. Prior to that, he worked as a systems engineer in the financial services sector with a focus on systems integration and replacement of dedicated communications networks and services for financial services over the Internet. Fredrik has a PhD in Industrial Engineering and Management with specialisation in Economic Information Systems and an MSc in Computer Science and Engineering.
Rate this talk
You must be logged-in to rate talks