A Former Minister's New Job: 'Regulatory Engagement'
Fenris Creations, the Reykjavik studio that makes EVE Online, announced on August 7, 2026 that Aslaug Arna Sigurbjornsdottir has joined as Senior Director, AI Partnerships. She previously served as Iceland's Minister of Justice and, before that, Minister of Higher Education, Science and Innovation. In the company's own announcement she described the move as an opportunity to help shape AI innovation by being part of a testing ground that enables what she called meaningful regulatory engagement - language that frames the job as policy work, not just business development.
Fenris CEO Hilmar Veigar Petursson framed the hire around the studio's newest project, EVE Frontier, a spinoff MMO where players called Riders survive in a post-collapse galaxy. Persistent virtual worlds, he said, are the next frontier, and EVE Frontier gives the studio something that does not exist anywhere else: a living, player-driven economy old and complex enough to study rather than simulate. That framing only makes sense once you see what Fenris built that economy alongside - a live AI research program with an outside lab already holding a financial stake in the outcome.
Why an AI Lab Bought Equity Instead of Just a Research License
Fenris only exists as an independent company because of a deal struck months earlier. In May 2026 it completed a management buyout from its Korean parent, Pearl Abyss, worth 120 million dollars - 100 million in cash plus 20 million in token acquisition rights. In the same window, Google DeepMind took an undisclosed minority equity stake in the newly independent studio and opened a research partnership, reportedly the first time a frontier AI lab has taken equity in a game studio specifically to secure a live game as a research environment. DeepMind's Adrian Bolton has said EVE Online demands skills AI has not yet fully mastered, including long-term planning and continuous learning.
The distinction between equity and a licensing deal matters. A license buys a snapshot of data; equity buys durable, ongoing influence over an environment that keeps evolving with its player base, its economy and its rules. That is a different kind of asset, and it explains why Fenris's next hire was not a data scientist but a former government minister whose entire career has been built on managing exactly the kind of relationships a lab-owned research environment will eventually need with regulators.
The Real-Money Gap: An AI Testbed Currently Outside the AI Act
EVE Online's in-game currency, ISK, is not a closed-loop token. Players buy PLEX - a purchasable item, roughly 20 dollars for 500 units - and sell it on the in-game market for ISK, at official rates of about 13 to 14 dollars per billion ISK, with informal secondary markets pricing it lower still. That direct, sanctioned convertibility between real money and the currency an AI agent would earn, spend and hoard inside EVE Frontier is exactly the kind of real-world-financial-stakes detail that tends to trigger high-risk classification once a regulator looks closely at an AI system.
The timing is the part no source coverage flagged. Most of the EU AI Act's remaining provisions became binding for EU member states on August 2, 2026, five days before this hire. Iceland, as an EEA-EFTA state alongside Norway and Liechtenstein, has not yet incorporated the Act into the EEA Agreement - that requires a separate EEA Joint Committee decision, a process that took GDPR the better part of a decade to complete after EU adoption. Icelandic organizations remain bound only for AI systems they place on the EU market or that affect people inside EU member states directly. A research program run out of Reykjavik, studying agents inside a real-money economy, currently sits in exactly that gap.
The Generalizable Lesson: Political Capital as an AI R&D Line Item
Reading the hire only as reputation management misses the more useful pattern. Iceland, Norway and Liechtenstein all sit on the EU AI Board as observers even without binding domestic obligations yet, which means Fenris now has, in Sigurbjornsdottir, a direct line into exactly the rooms where a case like real-money AI research environments would eventually get classified - before the rules exist, not after enforcement starts. That is a materially different move than hiring a compliance officer to react to a finished regulation.
The generalizable warning extends past one studio. Any platform with a live, player-driven economy with real-money convertibility - marketplace-heavy multiplayer games, virtual worlds, even some fintech-adjacent apps - is a candidate for the same playbook: an AI lab buys equity for durable access, and the platform hires political capital rather than just legal counsel to manage the gap between what regulators have written and what a frontier lab is already doing. Regulators tend to close loopholes only after a high-profile case forces the question; this is the kind of case that does that.
Read next: EVE Online Just Open-Sourced Its 20-Year Engine | A US Bill Could Switch Off the AI You Depend On
