Rather than relying solely on short-lived simulations that end and reset, the aim is to create a place where autonomous agents can be observed in a real, open and evolving economy, alongside real people, with real consequences, within the kind of dynamic system the company's games are uniquely able to provide.
The EVE Universe is particularly well-suited to this work. For more than two decades, it has developed as a living virtual economy, encompassing trade, production chains, currencies, commodity markets, alliances, conflict, cooperation, betrayal and loss, shaped by hundreds of thousands of real participants and all the unpredictable behavior that comes with them. Unlike chess or Go, the world has no end state: mistakes stay made, debts stay owed, relationships persist, and reputation follows you. That makes it one of the very few places where the behavior of autonomous systems can be studied under conditions that resemble a real society and a real economy, with all the irrational, emotional, and unpredictable reactions that real people bring with them.
In EVE Frontier, the new chapter of the EVE universe, autonomous AI systems will be studied within EVE Frontier’s live, open-ended world as they set goals, respond to change, and make coherent sequences of decisions over extended periods. This enables researchers to better assess model behavior in genuinely complex environments and inform the responsible development of future AI systems.
If the initiative succeeds, it opens an entirely new role for Fenris Creations and EVE Frontier, potentially making important contributions to research on intelligent agents, complex systems, and AI alignment.
What makes the work possible is the scale and longevity of what already exists. Fenris brings its experience in creating and operating a persistent virtual world, together with the engineering, design, and community expertise required to explore how autonomous AI behaves in a live, player-driven setting. Few virtual environments offer the same combination of persistence, complexity, and sustained human participation.
“Every major advance in AI has required a new proving ground. We believe persistent virtual worlds are the next frontier. EVE Frontier gives us something that doesn’t exist anywhere else: a living economy where intelligent agents can learn, cooperate, compete and make mistakes that have lasting consequences. We’re incredibly excited to welcome Áslaug to Fenris to help build the global partnerships needed to turn that vision into reality.” - Hilmar Veigar Pétursson, CEO
“Having seen firsthand how policy is developed and implemented, I am excited to bring my political experience to this new role. There is no greater opportunity to help shape the future of AI innovation than being part of a testing ground that empowers our partners to make meaningful inroads with regulators, helping ensure safe, practical and reliable AI standards. I look forward to helping Fenris take this work forward responsibly.” - Áslaug Arna Sigurbjörnsdóttir, Senior Director, AI Partnerships
Earlier this year, Fenris Creations announced a research partnership with Google DeepMind focused on exploring intelligence in complex, dynamic systems. That collaboration is already active, with teams from both organizations completing a two-day workshop this week to exchange ideas and identify practical areas for further work. The sessions reflected a strong alignment in approach and a shared interest in how persistent virtual worlds can support meaningful research and new forms of interactive experience. Áslaug’s appointment will broaden this work by helping Fenris Creations build relationships across the wider AI research, academic, policy and strategic landscape, bringing new institutions, perspectives and opportunities into the initiative as it develops internationally.
