The politics of artificial intelligence (AI) is set to drastically change in 2026 as recent technical breakthroughs get implemented across the economy. The public are understandably worried about AI and, so far, governments have struggled to articulate a clear vision for what it would mean for AI to go well.

Governments must stand ready to both protect people from the risks of AI and deliberately steer this transformation towards public value. But policy has, so far, been too timid to do so.

This report draws reflections from work so far on AI policy and highlights next steps, with recommendations for European governments seeking to demonstrate that they are intervening ambitiously in their citizens’ interests. It also introduces a how-to guide for directing AI to public value, identifying priority policies for the near term.

Rather than choosing between uncritical 'AI acceleration' and outright resistance, the report calls for a new approach: 'AI directionism' – where government actively steers the development and deployment of AI towards clear public benefits.

Key findings

• AI could deepen inequality, concentrate power and disrupt jobs without intervention.
• There is an urgent need for government to spread the benefits as industry grows.
• Current United Kingdom government policy is too narrowly focused on accelerating AI growth.

Recommendations

• Redistributing windfall gains from Sovereign AI investments back to the public.
• Deploying AI engineers to schools, hospitals and local government to experiment with where AI can improve outcomes.
• Reforming tax and subsidy schemes, so firms are rewarded for raising worker productivity rather than automating roles.
• Strengthening competition enforcement to prevent excessive concentration of power in the AI economy.