The European Central Bank (ECB) has warned that the dangers AI poses to the eurozone continue to grow ahead of its latest financial stability review.
In his keynote speech at Euro Finance Week Monday, ECB vice president Boris Vujčić told the audience that AI-related risks pose a major threat to financial stability. In last year’s assessment, he said, AI-related vulnerabilities were primarily around the risk that market sentiment could shift suddenly, collapsing the “stretched valuations” of AI stocks.
Now, AI is both more ingrained in the global financial system and more advanced, meaning it poses a new set of threats: the expansion of external funding from equity to debt markets, concentration of exposures, quickening AI adoption and potential cyberattacks are all major ECB concerns, Vujčić said.
The scale of financing required by firms along the AI supply chain is “unprecedented,” he continued. AI-linked debt issuance is at record highs, and is likely to continue to rise as cumulative capital expenditure by major US technology firms is projected to reach $5.5 trillion by 2030.
Echoing concerns raised by US financial firms in a Federal Reserve report, Vujčić also noted the growing trend of using “opaque” off-balance-sheet and circular financing arrangements to fund AI infrastructure development.
IPO volumes are now on track to exceed the dot-com bubble high, relative to US GDP, he added, noting that these moments have often coincided with market peaks, pointing to the “risk of excessive investor optimism”.
Additionally, AI cyber risks, and relatedly the high concentration of third-party providers of critical systems, are enough of a concern that the ECB published a special financial stability review feature on the topic prior to the speech.
The document, decoding the cyber-AI nexus, outlines the “significant implications” for financial stability in the euro area posed by the rapid development of AI’s cyber capabilities.
Advanced models are “enabling attackers and defenders alike to enhance their capabilities,” the authors say, with highly capable and agentic AI systems increasing the speed and sophistication of attacks while also introducing new vulnerabilities.
While both sides are gaining new AI-powered tools, they could end up favouring attackers, the paper warns, as low-cost AI agents lower entry barriers and defenders remain constrained by the cost of deploying frontier models, governance challenges and regulation-bound decision-making.
Mitigating these risks will require “banks, AI developers, cybersecurity firms and regulators to coordinate their efforts to enhance resilience, improve governance and address emerging systemic vulnerabilities,” it says.










Recent Stories