AI agents can analyze and iterate far faster than humans, potentially outpacing traditional regulatory capacity, making dedicated time, thoughtfulness, and resource allocation critical to financial stability.

There is a push for containment through circuit breakers and ‘kill switches’ to pause AI-driven market activity when needed, with openness to testing and governance mechanisms to prevent feedback loops and systemic stress.

Regulators are in early stages, forming consortia and collaborating with firms, hyperscalers, model developers, and academics to understand uses and risks, while stressing readiness for tech surprises.

Historically, rules can produce unintended responses, and AI expands how agents may evade or exploit regulation.

Policymakers should consider simpler, robust rules, such as higher capital requirements, and balance transparency to avoid gaming while preserving innovation.

The 10-year outlook should measure AI-enabled productivity, growth, trust, and stability rather than regulation volume; success hinges on nimbleness and broad adoption with safeguards.

Sarah Breeden of the Bank of England discusses AI regulation, financial stability, and collaboration with regulators and industry.

Regulators may regulate AI with AI themselves, but human judgment and supervisory oversight remain essential during the transition.

Closing note: substantial work ahead and the value of Wharton’s collaboration with regulators and the financial system to address AI’s role in finance.

AI is reshaping financial markets and policy challenges, prompting questions about how regulation should respond to technological shocks and behavioral adaptation.

Agentic AI raises stakes as autonomous systems pursue objectives, challenging whether current regulation can handle speed and scale without human oversight.

Humans-in-the-loop governance is essential, focusing on system design, testing, monitoring, guardrails, accountability, and explainability in complex AI models.