# Bank of England’s Breeden questions adequacy of AI regulations, proposes kill switches for financial markets

> Source: <https://cryptobriefing.com/boe-breeden-ai-regulation-kill-switches/>
> Published: 2026-09-21 15:16:14+00:00

# Bank of England’s Breeden questions adequacy of AI regulations, proposes kill switches for financial markets

The BoE deputy governor warned that autonomous AI agents could trigger financial meltdowns and called for circuit breakers to halt trading during catastrophic failures

Sarah Breeden, the Bank of England’s Deputy Governor for Financial Stability, told an audience of central bankers and policymakers that existing regulations may no longer be fit for purpose when it comes to artificial intelligence in finance. Speaking at the European Central Bank’s annual forum in Sintra, Portugal on June 30, Breeden specifically targeted “agentic AI,” the class of autonomous systems that make decisions and execute actions without a human in the loop.

Her proposed solution is about as dramatic as regulatory language gets: market-wide circuit breakers and kill switches capable of shutting down trading entirely if an AI system goes haywire.

## The problem with robots that don’t ask permission

Agentic AI is different from the chatbot your bank uses to answer questions about overdraft fees. These systems autonomously execute trades, process payments, and make commercial decisions at speeds no human could match.

Breeden warned that correlated AI agents, meaning multiple autonomous systems trained on similar data and pursuing similar strategies, could amplify market volatility during periods of stress.

A survey from the Cambridge Centre for Alternative Finance found that roughly 52% of financial firms were already deploying agentic AI as of mid-2026. More than half the industry has adopted a technology that the deputy governor of one of the world’s most important central banks now says may not be adequately regulated.

Beyond market volatility, Breeden flagged systemic cyber threats tied to advanced AI models. She also raised thorny questions about consent and liability. When an AI agent autonomously decides to execute a trade that goes sideways, who’s responsible? Current legal frameworks weren’t designed to answer those questions cleanly.

### AI, tech, and the markets they move—in one daily briefing.

Daily. Free. Join 34,000+ readers across crypto, finance, and policy.

## Kill switches and crisis protocols

The most eye-catching proposal involves market-wide circuit breakers, essentially emergency stop buttons that regulators could activate to freeze trading if an AI-driven cascade starts spiraling out of control.

She also recommended enhanced recovery protocols that would allow one bank to assume another institution’s essential functions during an AI-related crisis.

These proposals represent a meaningful departure from the BoE’s prior position. Until recently, the central bank maintained that existing regulatory frameworks were sufficient to handle technological innovation in finance. Breeden’s address effectively retired that stance.

## A broader regulatory reckoning

The Bank of England has been collaborating with the Financial Conduct Authority and participating in a public-private AI Consortium to evaluate and strengthen the financial system’s resilience against emerging AI threats. Her Sintra speech appears to be the most public articulation yet of conclusions that have been building through those quieter channels.

For financial institutions, the implications are practical and immediate. Firms that have aggressively deployed agentic AI may face new compliance requirements that didn’t exist when they made those technology investments. The 52% of firms already using these systems will need to evaluate whether their current implementations can accommodate circuit breaker mechanisms, enhanced reporting obligations, and potentially new liability frameworks.

If regulators gain the authority to freeze markets when AI systems malfunction, that introduces a new category of risk for any strategy that depends on continuous market access. Algorithmic trading firms, high-frequency desks, and even passive index funds that rely on AI-driven rebalancing would need to account for the possibility of sudden, externally imposed trading halts.

**Disclosure:** This article was edited by Editorial Team. For more information on how we create and review content, see our

[Editorial Policy](https://cryptobriefing.com/editorial-policy/).
