28/05/2026
The article I wrote for the Digital Growth Collective on «Shadow AI and Democratic Accountability» has been published today on the DGC Insights Hub (London).
The piece examines a specific problem I have been studying: what happens when tools begin shaping institutional decisions without ever entering formal governance processes. Credit assessments informed by AI assistants, candidate screening through generative platforms, sensitive data processed through tools never authorized by the organization.
For boards, executives, and risk committees, the issue is not AI adoption. It is invisible decision-making at scale. Institutions may continue governing formally while progressively losing visibility over how decisions are actually being produced, by whom, and on what basis.
In Latin America and other high-informality economies, where institutional visibility is already unevenly distributed across formal and informal sectors, Shadow AI does not only create operational risk. It redistributes the question of who is accountable when decisions are influenced by systems no one formally governs.
The piece argues that the relevant governance question is no longer about regulating formal AI systems. It is about whether institutions can recognize how AI is actually being used inside their workflows, and whether they retain the ability to explain, attribute, and contest the outcomes those systems shape.
Thanks to Mike Flache and the Digital Growth Collective team for publishing this piece. I would welcome perspectives from boards, investors, and governance practitioners on how this is playing out inside your own organizations.
Read the full article on the DGC Insights Hub here: https://www.linkedin.com/pulse/shadow-ai-democratic-accountability-governance-jdo3e/
