Jack Clark, co-founder of the US-based AI lab Anthropic, has released a new study that sketches three divergent paths for the continent's economy as artificial intelligence matures. The research, published alongside an interactive modelling tool, lets users adjust assumptions about AI progress and corporate uptake to see how gross domestic product (GDP) and employment might evolve by the end of the decade.
What the three scenarios entail
In the first, AI acts as a supportive assistant for workers, delivering productivity gains comparable to the early internet. Growth is steady and broadly shared, mirroring historical patterns of technological diffusion.
The second scenario assumes that by 2030 AI can perform roughly half of all knowledge-intensive tasks autonomously. While overall economic output could double the usual rate, wages for knowledge workers would stagnate and the benefits would accrue mainly to other sectors.
The most extreme outlook envisions AI surpassing human capability in almost every knowledge-work activity, handling the bulk of tasks without human input. Under this model GDP would rise by about 15 % a year, effectively doubling the economy every four and a half years, but unemployment would surge to levels unseen in a typical recession.
Why the study matters for Europe
Europe faces a unique set of challenges and opportunities as AI spreads. The European Union's AI Act already imposes strict transparency and safety requirements on high-risk systems, which could shape the speed of adoption. Moreover, the continent's labour market is characterised by a higher proportion of public-sector and regulated jobs, meaning that any abrupt displacement could have pronounced social repercussions.
Anthropic's model focuses solely on the supply side, how much AI can boost productivity, and deliberately omits demand-side factors such as consumer spending, fiscal policy or potential financial-market shocks. The authors acknowledge that the framework is a "stark simplification of a complex reality". Nevertheless, the scenarios provide a concrete basis for economists and policymakers to discuss the trade-offs between growth, inequality and social stability.
What comes next
Clark stresses that the speed of AI diffusion is likely to be slower than many industry forecasts. In an interview with NPR he said,
"The technology itself will keep improving at a very, very fast and sustained rate, but it will spread through the economy more slowly than most people assume."
If adoption accelerates, the fiscal windfall from higher GDP could give governments room to fund retraining programmes and safety nets, a possibility that Clark describes as "unimaginable today". Conversely, a sluggish rollout would keep growth modest and keep the labour market relatively stable.
European leaders are already calling for more rigorous analysis. Earlier this year, over two hundred economists, including Nobel laureates and former tech CEOs, urged policymakers to commission systematic research on AI's macro-economic impact. Anthropic's interactive tool is a step towards meeting that demand, allowing analysts to test how different policy levers, such as tax incentives for AI-driven upskilling, might alter the projected outcomes.
In the short term, the data from business-spending trackers such as Ramp's AI Index suggest that corporate investment in AI is rising only marginally, with adoption rates inching up by a few tenths of a percentage point each month. This measured pace supports the view that Europe is likely to experience a gradual, internet-age-style transformation rather than a sudden explosion of wealth.
Stakeholders will need to monitor both the technological trajectory and the regulatory environment. As the EU refines its AI governance framework, the balance between encouraging innovation and protecting workers will shape which of the three futures, if any, becomes reality.