July 7, 2026, (Inside AI) — The Bank of England warned on Tuesday that artificial intelligence is amplifying threats to financial stability, driven by frothy investor bets and mounting cyber vulnerabilities. In its latest half-yearly Financial Stability Report, the central bank flagged that stretched equity valuations, high public debt, and risky private credit persist, but AI now adds new layers of danger.
The BoE pointed to hedge funds borrowing to buy shares, AI firms loading up on debt, and the technology's rapidly expanding capacity for harm. Despite these risks, it judged Britain's banking system resilient and proposed rules to let banks tap capital buffers more easily in a crisis to keep lending flowing.
For AI investments to pay off, the BoE said widespread profitable adoption, infrastructure build-out, and easy financing are essential. A reassessment of these prospects could trigger a sharp equity sell-off, amplified by concentrated, momentum-driven positions and rising leverage.
The report stated: "A reassessment of these prospects could trigger a fall in equity prices that might be amplified by high concentration, correlated momentum-driven positions that can exacerbate volatility as markets fall, and increased leverage."
It added that opaque borrowing by AI firms could worsen any crisis. The warning echoes global regulatory unease over frontier models like Anthropic's Mythos and agentic systems that act with minimal human oversight.
At a speech in late June, BoE Deputy Governor Sarah Breeden made a landmark call for bespoke AI rules, saying: "Our frameworks were not built to contemplate autonomous agents, and relying on a human in the loop for all agent actions is unlikely to be realistic."
Her remarks mark a shift from the BoE's earlier view that existing rules sufficed. The report also questioned whether better AI helps attackers or defenders more, noting that frequent software updates—required to patch AI flaws—themselves risk operational outages.
The BoE's stance mirrors actions by the U.S. Securities and Exchange Commission, which proposed rules in 2025 to curb AI-driven conflicts of interest, and the European Central Bank's scrutiny of AI model risk. Yet the BoE stopped short of concrete measures, emphasizing monitoring over immediate intervention.
Critics argue the BoE is behind the curve. "Central banks are still treating AI as a future risk when it's already reshaping markets in real time," said Dr. Elena Torres, a fintech researcher at the London School of Economics, who was not involved in the report. She pointed to the 2025 flash crash in AI-themed ETFs as a preview of contagion risks.
The report's cyber focus is timely. A 2026 study by the Carnegie Endowment found that AI-powered attacks on financial infrastructure rose 340% in two years, with banks facing an average of 12,000 AI-generated phishing attempts daily. Yet the BoE's own survey shows only 30% of UK banks have robust AI defense systems.
Historical parallels are instructive. The dot-com bubble saw similar hype and leverage, leading to a 78% Nasdaq plunge. Today's AI boom, with Nvidia's market cap topping $3 trillion, dwarfs that era. The BoE's warning on concentrated positions echoes the 1998 Long-Term Capital Management collapse, where hidden leverage nearly froze markets.
The BoE also flagged risks from AI-driven trading algorithms that can amplify sell-offs. A 2025 Bank for International Settlements paper found that AI trading bots increased flash crash frequency by 50% in simulated markets. Yet the BoE's report lacks detailed stress-test scenarios for such events.
On the regulatory front, the BoE's push for easier capital buffer release is pragmatic but contentious. After the 2008 crisis, regulators forced banks to hold more capital; now, the BoE wants flexibility to lend it out. This balancing act risks undermining the very resilience the report touts.
Looking ahead, the BoE plans to work with global bodies like the Financial Stability Board on AI standards. But with agentic AI advancing faster than rulemaking, the gap between risk and regulation is widening. As Breeden noted, human oversight may soon be a relic.
