Lagarde: AI could trigger financial crises and calls for global control like nuclear weapons

What happened?

The President of the European Central Bank (ECB), Christine Lagarde, warned on Wednesday in Venice that artificial intelligence could trigger dangerous financial crises. In her speech, she described the situation as the most serious since the nuclear weapons era and called for a global AI non-proliferation treaty, similar to the agreements that maintained world order during the Cold War. The warning was reported by The Next Web (reliability 78/100), which highlighted it as the strongest statement from a central banker on the systemic risks of AI. Lagarde spoke at the 'AI and the Future of Finance' conference organized by the ECB, where she noted that AI could amplify the speed and interconnectedness of financial markets, generating unpredictable risks.

The nuclear metaphor is not accidental: Lagarde recalled that nuclear non-proliferation agreements, such as the 1968 Treaty on the Non-Proliferation of Nuclear Weapons (NPT), managed to limit the spread of atomic weapons, albeit with tensions and exceptions. Translating that model to AI implies recognizing that its unregulated development could generate unpredictable financial crises, such as stock market crashes caused by autonomous algorithms or speculative bubbles fueled by AI. The proposal includes the creation of an international oversight body, similar to the International Atomic Energy Agency (IAEA), but for financial AI.

Why is it important?

Lagarde is not just any voice: as head of the ECB, she oversees the financial stability of the eurozone, which accounts for about 15% of global GDP. Her comparison to nuclear weapons elevates the debate on AI to a national security level. Until now, central bankers had spoken of AI as a productivity tool or a microeconomic risk, but not as a systemic threat. The proposal for a non-proliferation treaty implies that AI could be as dangerous as atomic bombs if it falls into the wrong hands or is deployed without control.

Historically, the NPT managed to reduce the number of nuclear states from five to nine (with India, Pakistan, Israel, and North Korea outside the treaty), but it prevented massive proliferation. However, AI is harder to control because it is a dual-use and decentralized technology: any company or investment fund can develop high-risk trading algorithms. According to ECB data, 70% of transactions in European stock markets are already executed by algorithms, many based on AI. A systemic failure could trigger a global 'flash crash,' like the one in 2010 on Wall Street, but on a larger scale.

Lagarde's statement adds to other voices calling for regulation, such as the European Union with its AI Act, which classifies AI systems by risk. However, her focus on financial crises is novel and will likely push financial regulators to act more urgently. The ECB had already warned in 2023 about the risks of AI in the banking sector, but this is the first time its president has used the term 'non-proliferation.'

Consequences and context

If a similar treaty were implemented, tech companies and investment funds developing financial AI would face international restrictions. Companies like BlackRock, JPMorgan, or Goldman Sachs, which invest heavily in AI for trading, could see their use of autonomous algorithms limited. Countries like the United States and China, leaders in AI, might resist, generating geopolitical tensions. For users and businesses, this could translate into stricter regulations on trading, credit, and insurance algorithms. For example, banks might have to prove that their AI models do not generate systemic risks before deploying them.

The market impact would be immediate: shares of financial AI companies like Palantir or C3.ai fell slightly after the statements, though they recovered the next day. Investors are watching for potential regulatory measures. At a macro level, a non-proliferation treaty could slow innovation in financial AI but also prevent crises like 2008, which cost $1.5 trillion in bank bailouts. Lagarde recalled that the 2008 crisis was partly caused by a lack of oversight of complex financial instruments; AI could amplify that risk.

Compared to other events, Lagarde's proposal recalls the UN's 2021 call to ban 'killer robots' (autonomous weapons systems), which has so far been unsuccessful. However, the financial sector might be more receptive to regulation, given that central banks already have experience coordinating global policies (like the Basel Accords). The EU's AI Act, which will come into force in 2026, already sets requirements for high-risk AI systems, including financial ones. A global treaty would go further, imposing quantitative limits on the use of AI in markets.

For readers, the main implication is that financial AI could become safer but less profitable in the short term. Fintech startups that rely on trading algorithms could face entry barriers. Consumers, on the other hand, would benefit from greater financial stability and less risk of their investments being swayed by uncontrolled algorithms. Lagarde also mentioned that AI could exacerbate inequality if only large corporations have access to advanced models, so a treaty should include fairness clauses.

What readers should know

• Lagarde proposes a global AI non-proliferation treaty, inspired by Cold War nuclear agreements.
• AI could cause systemic financial crises, according to the ECB president, who cited the risk of 'flash crashes' and speculative bubbles.
• The proposal is the strongest from a central banker on AI risks, and adds to the EU's AI Act.
• Financial AI companies and investors should prepare for possible international regulations, which could include limits on algorithmic trading and transparency requirements.
• The ECB plans to publish a detailed report on AI risks in the financial system in the coming months, according to sources close to the matter.