While politicians argue over which regulations might slow AI down, the real question is whether AI can save growth itself — or whether we're automating our way into stagnation with better graphics.
The Summary
- Exponential View's latest explores AI's political moment as governments worldwide scramble to regulate systems they barely understand, while researchers debate whether AI will actually deliver the productivity gains everyone's betting on
- The newsletter highlights a widening gap between AI's technical capabilities and its measurable economic impact — the productivity paradox might be back, just with better chatbots this time
- Key tension: regulatory frameworks are crystallizing before we know if AI will be the growth engine of the 2030s or an expensive detour from addressing real bottlenecks in energy, housing, and human capital
The Signal
AI politics is heating up across major economies, but the regulatory debates miss something crucial. We're legislating the guardrails for a technology whose economic impact remains surprisingly unclear. Yes, AI can write code and generate images. But can it actually reverse declining productivity growth in developed economies? That question is wide open.
The piece examines how political systems are responding to AI's rapid development, from the EU's comprehensive regulatory approach to the US's more fragmented framework. But there's a disconnect. Policymakers are treating AI as an inevitable force that needs containing, while economists are still trying to figure out where the productivity gains actually show up.
"We're building regulatory architecture for AI's promise, not its proven track record."
The broader context matters here: advanced economies have struggled with productivity growth for decades. Real wages have stagnated. Housing costs have exploded. Infrastructure ages. If AI is the solution to this growth crisis, we should be seeing early signals by now — companies that adopted AI aggressively should be dramatically outperforming peers. Some are. Most aren't.
The newsletter also touches on energy implications, which connects directly to AI's viability at scale. Training large models requires enormous power. Inference costs are dropping, but total energy demand keeps climbing as usage expands. If AI is supposed to solve growth problems, it can't simultaneously create an energy crisis.
Key questions emerging:
- Can AI generate real productivity gains before its energy demands strain grid infrastructure?
- Are current AI capabilities actually bottlenecked by compute, or by the messy problems of implementation in real organizations?
- Will AI regulation lock in current market leaders (OpenAI, Anthropic, Google) or create space for new approaches?
The political moment for AI is arriving just as the economic case remains unproven. That timing creates risk. Bad regulation based on hype rather than measured impact could either kill genuinely useful applications or entrench incumbents who haven't yet delivered on their promises. Neither outcome serves the goal of sustainable growth.
The Implication
Watch the gap between AI capability and AI impact. If that gap doesn't narrow substantially in the next 18 months — if companies keep spending billions on AI infrastructure without corresponding gains in output per worker — the political conversation will shift from "how do we regulate this transformative technology" to "why did we bet so much on this." That shift would reshape funding, talent allocation, and which AI applications actually get built. The growth question matters more than the politics.