At FOSDEM 2026, Michiel Leenaars will explore how geopolitical conflict, resource scarcity, and AI-generated code threaten the foundations of open source collaboration, proposing defensive strategies for preservation.

The free and open source software ecosystem faces unprecedented challenges as geopolitical tensions, resource constraints, and generative AI technologies converge to threaten its collaborative foundations. At FOSDEM 2026, Michiel Leenaars will present "FOSS in times of war, scarcity and (adversarial) AI," examining critical vulnerabilities in today's digital landscape and proposing defensive countermeasures.
Open source software emerged during a period of relative geopolitical stability following the Cold War, enabling global cooperation that created unprecedented public digital goods. This environment fostered innovation where users could actively shape technology rather than passively consume it. However, Leenaars argues this era has ended abruptly. Nation-states now weaponize open source infrastructure for disinformation campaigns and digital warfare, while authoritarian regimes exploit FOSS to build surveillance architectures. The Snowden revelations exposed how even democratic governments maintain contradictory agendas regarding digital freedoms.
Compounding these threats, climate change and resource scarcity create material constraints on technology's physical foundations. As global populations consume more energy and raw materials, the environmental cost of computation becomes increasingly unsustainable. Meanwhile, generative AI introduces new vulnerabilities at the code level. Large language models promising developer efficiency instead create opaque supply chain risks. These systems lack ethical frameworks and can introduce subtle vulnerabilities through manipulated training data - what Leenaars terms "adversarial AI."
The core danger lies in AI's ability to rewrite truth from within, creating phantom code that undermines decades of carefully engineered trust. Unlike identifiable binary blobs, AI-generated code operates as a black box where malicious triggers can be embedded during training and remain undetected until activation. This threat multiplies when combined with wartime incentives to compromise software integrity.
Leenaars proposes a framework of "maximal defendable FOSS" combining compartmentalization, formal verification methods, and hybrid human-AI workflows. This approach maintains human oversight while implementing cryptographic proofs and symbolic analysis to verify system integrity. The solution acknowledges AI's potential while rejecting blind trust in opaque systems, emphasizing that FOSS's survival depends on adapting verification mechanisms to counter new threat vectors.
The talk challenges the community to confront uncomfortable realities: that technologies built for empowerment can be weaponized, that environmental limits impact digital infrastructure, and that automation without transparency risks the ecosystem's foundations. Rather than retreating, Leenaars frames this as an opportunity to reinvent collaborative safeguards for a more hostile digital age.
FOSDEM 2026 takes place January 31-February 1 in Brussels. Michiel Leenaars' session occurs Saturday at 10:00 in the Janson auditorium. Conference schedule | Main Track details

Comments
Please log in or register to join the discussion