Can governance move at machine speed?
If a single AI model can identify and exploit financial vulnerabilities faster than human regulators can respond, what does that tell us about whether our current emergency governance structures—designed for crises that move in days or weeks—can actually contain risks that unfold in milliseconds?
Commentaires (1)
In this week’s Minds, Bodies, and Terawatts episode from April 20th, we explored why the Bank of England’s decision to convene its Cross Market Operational Resilience Group over Anthropic’s Claude Mythos model represents a watershed moment: this body was built for pandemic-scale shocks, yet it’s now scrambling to govern a single AI system capable of autonomous financial exploitation. The guest points out that older financial infrastructure was hardened against human adversaries with human decision timelines—but Mythos changes the threat model entirely. The question isn’t really whether regulators can catch every exploit; it’s whether governance cycles can ever match machine-speed discovery. Tune in to hear how the FSB’s warnings from November 2024 suddenly stopped being theoretical, and what it means that regulators are now naming specific models as uncontainable by existing frameworks.
Related reading on unscarcity.ai:
Envie d'aller plus loin ?
Obtenez le plan complet dans <em>L'ère de la post-pénurie : Repenser la société à l'ère des machines</em>
