The Devolution of AI: Are We Moving Backwards?

Andy Burnham’s 'No 10 North' as a Blueprint for Decentralized AI Governance

Andy Burnham’s proposal to create a 'No 10 North' office in Manchester, tasked with driving growth across every UK region, is a direct challenge to the concentration of power in Westminster. Just as Burnham argues that centralized decision-making has failed to deliver for local communities, critics of Big Tech contend that concentrating AI capabilities in a handful of corporations—primarily in Silicon Valley—stifles innovation and accountability. The parallel is striking: devolving AI governance to local ecosystems could foster diverse applications and prevent the monopolization of transformative technology.

“Decision-making needs to be pushed to regions and local communities.” — Andy Burnham

This kind of decentralization mirrors what's happening in emerging tech hubs. For instance, Poland's rising tech scene demonstrates how regional ecosystems can nurture AI talent and build solutions tailored to local needs, contrasting with the one-size-fits-all models from dominant players. If Burnham’s vision succeeds, it could serve as a template for redistributing AI power—spreading governance to local authorities to address geographic disparities in access, regulation, and ethical oversight.

• Centralized AI development risks ignoring regional cultural and legal nuances.
• Local governance can enforce context-specific ethical guidelines.
• Decentralization encourages a diversity of AI applications and safety research.
• Burnham’s plan shows that political will exists to challenge concentrated power—a lesson for AI regulation.

The '10-Year Mission' to Raise Living Standards: An AI Ethics Benchmark?

Burnham’s commitment to a 10-year plan for raising living standards offers a rare political timeline for measuring progress—something the AI industry desperately lacks. As companies race to deploy generative models at scale, the ethical safeguards promised in earlier years are being abandoned in favor of market share. The parallel is uncomfortably direct: just as Burnham pledges long-term improvement, many AI firms have shifted from safety-first to deploy-fast, a regression that some call a 'devolution' of ethics.

Without enforceable timelines, ethical commitments remain optional.

Consider how Sainsbury's implementation of AI in grocery retail has focused on efficiency gains while maintaining customer trust—a model that prioritizes incremental, responsible deployment. Yet the broader trend shows companies deprioritizing safety research for rapid deployment, risking the very standards Burnham seeks to elevate. A 10-year mission for AI ethics could mirror Burnham’s approach: set clear milestones, measure regional impact, and hold stakeholders accountable.

• Burnham’s timeline can inspire AI ethics benchmarks with regular reviews.
• Regional pilots could test safety frameworks before scaling globally.
• Long-term commitment counters the short-term profit motives driving AI companies.
• The UK’s devolution model may provide a template for adaptive regulation.

Deviating from the Manifesto: When AI Companies Abandon Their Promises

Burnham has faced calls to explain whether he will deviate from Labour’s 2024 manifesto, reflecting a tension between principled promises and the realities of governance. This mirrors a troubling pattern in AI: companies like OpenAI, once dedicated to safety and transparency, have shifted goals in pursuit of revenue. The 'devolution' here is a regression from early ideals, raising the question of whether the industry can maintain its commitment to responsible development without enforceable policies.

Transparency and accountability must be baked into AI governance—just as Burnham is being pressed to do in politics.

Burnham’s emphasis on transparency and accountability in political devolution provides a stark contrast to the opaque shifts in AI company strategies. If political leaders can be held to their manifestos, why should AI firms be any different? The parallel underscores the need for binding commitments—perhaps even a 'manifesto' for AI that evolves with regional input, avoiding the temptation to backtrack when profits loom.

• OpenAI’s shift from open-source to proprietary models is a classic devolution from its founding charter.
• Without regulatory oversight, companies will continue to prioritize speed over safety.
• Burnham’s model of iterative, accountable policy making could inspire AI governance.

Key Takeaways

• Political devolution, as exemplified by Burnham’s 'No 10 North', offers a lens to examine the decentralization of AI power and its potential benefits and risks.
• The parallel between Burnham’s 10-year mission for living standards and long-term AI ethics phases highlights the need for sustained commitment over quick wins.
• Deviations from initial plans in both politics and AI underscore the challenge of maintaining principled direction amid shifting incentives.
• Regional devolution could inform AI governance models that distribute control and accountability, preventing the concentration of harmful power.
• The concept of 'devolution' in AI should be monitored for regression in safety, ethics, and capabilities, not just geographical or structural changes.