AI as Economic Architect, Not Just a Tool

Philip R. Lane, Member of the European Central Bank's (ECB) Executive Board, officially designated artificial intelligence (AI) as a general-purpose technology (GPT) on par with electricity and the internet, delivering an in-depth analysis of its macroeconomic implications for the euro area. Speaking at the ECB-SAFE-RCEA International Conference, Lane emphasized that AI's potential lies not in any single application but in its capacity to reshape entire production processes, business models, and economic structures.

Lane pinpointed what fundamentally distinguishes AI from previous GPTs. While steam power, electrification, and ICT made existing processes faster and cheaper, AI has the potential to raise the productivity of the innovation process itself — accelerating scientific discovery, shortening R&D cycles, and compressing the time from knowledge creation to commercial application. In essence, AI may not just shift the level of productive capacity, but the rate at which that capacity grows.

Why the ECB Is Talking About AI

When a central bank formally addresses AI economics on a public stage, it signals more than intellectual curiosity. Productivity gains directly affect potential growth rates, which in turn shift the baseline for interest rates and monetary policy. The ECB's elevation of AI to a core agenda item reads as a signal that the monetary policy framework itself is under reassessment.

Lane also highlighted the possibility of computing power ("compute") becoming a primary input into the innovation process, simultaneously accelerating progress across multiple frontier technologies. If realized, this could structurally raise potential growth rates — with direct implications for long-run neutral interest rate estimates.

The Forecast Gap: Why Such Extreme Divergence?

Estimates of AI's economic impact diverge dramatically. The range Lane himself cited spans an extraordinary breadth.

Goldman Sachs projected in March 2023 that widespread AI adoption could drive a 7% increase in global GDP over a decade, raising annual labor productivity growth by around 1.5 percentage points. McKinsey went further that June, suggesting AI combined with broader automation could add up to 3.4 percentage points per year to productivity growth through 2040.

More recent estimates tell a different story. Acemoglu (2025) concludes that total factor productivity (TFP) gains over the next decade are unlikely to exceed 0.66% in total — a marginal annual contribution. The OECD projects 0.4 to 1.3 percentage points of additional annual labor productivity growth for high-AI-exposure countries like the US and UK. For the euro area specifically, Bergaud (2024) estimates an annual productivity boost of just 0.29%.

This divergence is not simply a matter of optimism versus pessimism. The core variable is how many tasks and industries prove resistant to automation — what Lane termed the "weakest link" constraint. If significant portions of the economy cannot be automated, aggregate AI impact shrinks considerably.

The Emerging Frontier: Agentic AI

The technological trajectory itself is shifting rapidly. AI has progressed from narrow machine-learning pattern recognition to large language models (LLMs) and generative AI platforms, and now "agentic AI" is emerging as the next frontier. Agentic AI may act as an independent economic agent rather than merely augmenting human effort — setting goals and executing tasks without human direction. If this materializes, labor markets and production organization could face fundamental restructuring.

Outlook [AI Analysis]

As the ECB formalizes its AI economic analysis, AI-driven productivity shocks are likely to become a major variable in future monetary policy discussions. Upward revisions to potential growth rates would affect long-run neutral interest rate estimates, with ripple effects across the policy frameworks of the ECB, the US Federal Reserve, and other major central banks.

In scenarios where compute becomes a primary innovation input and AI accelerates frontier technologies across the board, potential growth rates could structurally increase. However, if automation-resistant industries constitute a significant share of the economy, macroeconomic effects are likely to remain contained.

For the euro area, the balance between digital infrastructure investment and the EU AI Act's regulatory framework is likely to prove a critical determinant of AI adoption speed. Lane's speech signals that AI has moved beyond technology policy — it has officially entered the center of the macroeconomic policy agenda.