The European Central Bank is now integrating artificial intelligence into the heart of its monetary analysis processes. A technical evolution presented as progress by its proponents, but one that worsens an already structural democratic deficit - and raises fundamental questions about the legibility of monetary power in Europe.
Algorithms at the Service of Interest Rate Policy
On July 6, 2026, Philip Lane, Chief Economist of the European Central Bank, publicly outlined in an official speech how artificial intelligence is now integrated into the conduct of the eurozone’s monetary policy. The address, delivered within the ECB’s usual institutional framework, marks a milestone: the use of AI is no longer a discreet experiment; it is openly embraced and presented as a fully-fledged steering tool.
The applications in question fall into several categories:
- Predictive models of inflation and economic conditions, designed to anticipate price developments;
- Automated processing of massive data from financial markets, supply chains, and consumer behavior;
- Decision support on interest rate guidelines, via systems capable of aggregating signals that traditional econometric models failed to capture.
The ECB is not alone in this approach. On April 2, 2026, Denis Beau, First Deputy Governor of the Banque de France and member of the Eurosystem, declared to students in Brest that the institution is “closely monitoring all developments likely to influence economic conditions in the coming years as well as prices”, and that in France, investment in software and databases has doubled over the last decade - and multiplied by 2.5 in the United States over the same period.
An Institution Already Beyond Electoral Control - AI as an Additional Layer of Opacity
The ECB is not elected. Its Executive Board members are appointed by the European Council, on the recommendation of the Council of the EU, after consulting the European Parliament - a procedure that confers no direct popular mandate. This architecture, inherited from the Maastricht Treaty, was designed to protect monetary policy from short-term political pressures. It has also produced an institution whose decisions affect the borrowing rates of 340 millions citizens in the eurozone without them having any real democratic leverage.
The introduction of AI into this process adds an additional layer of illegibility:
- The predictive models deployed are, by nature, partial black boxes: their parameters, training data, and internal weightings are not subject to any comprehensive publication;
- Accountability for an interest rate decision becomes difficult to assign when it results from an algorithmic recommendation;
- Parliamentary oversight - already limited to periodic hearings - is even less equipped to question the logic of a machine learning system than to evaluate a classic econometric model.
“The rapid adoption of AI requires central banks to embrace this new technology,” declared the Bank for International Settlements (BIS) in its Annual Economic Report 2024 - without the question of democratic control over these tools occupying a central place.
The BIS itself acknowledges in the same report that “the work of central banks as guardians of the economy will be directly affected as frontline users of AI tools.” The admission is significant: monetary institutions are no longer content to analyze AI as an external economic phenomenon - they are submitting to it themselves as operators.
The Macroeconomic Impact of AI: What Central Banks Anticipate
Behind the question of tools lies that of systemic effects. European central banks are closely monitoring projections of AI’s impact on productivity and growth - as these variables directly condition their interest rate decisions.
The figures available in institutional sources are as follows:
- According to Cerutti et al. (2025), cited by the Banque de France, global growth could be increased by 0.1 to 0.4 percentage points per year thanks to AI, due to a productivity increase of 0.1 to 0.2 percentage points per year;
- According to Aghion and Bunel (2024), also cited, the adoption of AI over the next decade could increase annual productivity growth in developed countries by 0.1 to 1.2 percentage points per year, with a median estimate of around 0.7 points;
- The OECD estimates a range of +0.3 to +1.0 points of annual productivity for France;
- The BIS also highlights that “with widespread adoption, AI could reinforce firms’ ability to adjust prices more quickly”, with direct repercussions on inflation dynamics.
Sources: Banque de France, speech by Denis Beau, April 2, 2026, citing Cerutti et al. (2025), Aghion and Bunel (2024), and OECD estimates for France.
These wide ranges reveal considerable uncertainty - precisely what algorithms are supposed to reduce. It is within this space of uncertainty that AI is establishing itself as a decision-making tool, without European citizens having been consulted on the appropriateness of this delegation.
Institutional Convergence: ECB, BIS, and the Digital Transformation Agenda
The integration of AI into central banks is not developing in a vacuum. It is part of a digital transformation agenda for finance whose broad outlines are coordinated on an international scale - through the BIS, the IMF, and global economic governance forums.
The BIS, in its annual report 2024, explicitly states that “the increased importance of data as the cornerstone of the AI revolution accelerates the need for cooperation among central banks”. This enhanced cooperation implies a growing harmonization of tools, standards, and, implicitly, monetary policy directions - beyond national borders and local democratic mandates.
Several observations are worth noting:
- The IMF is cited by the Banque de France as a reference source for analyzing the effects of AI on investment: since early 2024, the dynamism of investment in advanced G20 economies is said to be mainly driven by the US semiconductor and software sector;
- In France, investment in data center construction has multiplied by 2.5 over the last decade, according to data cited by Denis Beau;
- The BIS identifies the financial sector as “one of the most exposed to the benefits and risks of AI,” with risks including more sophisticated cyberattacks - a vector of systemic vulnerability rarely mentioned in institutional communications.
The question of who designs and audits the algorithms used by the ECB remains, at this stage, without a satisfactory public answer. Major monetary institutions do not publish the list of their technology providers, nor the conditions under which private actors - primarily American tech companies - participate in the architecture of these tools.
Central Bank Digital Currency: The Logical Horizon
The use of AI in monetary policy cannot be dissociated from a broader project: that of central bank digital currency (CBDC), which the ECB is actively developing under the name of the digital euro. This project represents the logical culmination of the described trajectory: a currency entirely administered by an unelected institution, whose flows could be analyzed, directed, and conditioned by algorithmic systems.
The BIS notes that AI could improve payment systems - one of the direct application grounds for a CBDC. The convergence between the integration of AI into monetary analysis tools and the deployment of a sovereign digital currency outlines an unprecedented architecture of monetary power: centralized, automated, and fundamentally removed from ordinary democratic debate.
Conclusion
The integration of artificial intelligence into the conduct of European monetary policy is not, in itself, a technical anomaly. Central banks have always used complex quantitative models. What is changing is the scale of the opacity, the speed of the delegation, and the complete absence of a democratic framework to debate it.
When an unelected institution entrusts part of its decision-making processes to systems whose designers, parameters, and internal logic remain largely unknown to the public - and to the majority of European parliamentarians - the question that arises is not technical. It is constitutional: who governs, in whose name, and under what rules? The answer outlined by the ECB’s current trajectory deserves to be asked aloud, before the architecture is permanently in place.
