Politics and Policy

The Algorithmic State: Geopolitics and Governance in the AI Era

The development and governance of AI have transcended technological innovation to become a primary arena for great power competition. The contest for AI supremacy is not merely a race for technological advantage but a fundamental struggle to define the economic, social, and political operating systems of the 21st century. Wherever there’s powerful technology, there’s inevitably regulation – and AI represents the most significant regulatory challenge of our time.

This section, like others, is a high-level overview of some of the considerations being made.

Read more below. Relevant links in the footnotes (‘References’), although NB some are behind paywalls.

The Geopolitical Landscape: A World Divided

The global governance of AI is fragmenting into distinct, competing ideological blocs. National strategies reflect deeper geopolitical objectives, with major powers championing different models that mirror their broader visions for world order.

The American Model: Deregulation for Dominance

Market-Driven Approach: The US pursues “permissionless innovation” – the philosophy that drove internet growth now applied to AI. America’s AI Action Plan (July 2025) explicitly targets “unquestioned global technological dominance” while the Trump administration rescinded previous “Trustworthy AI” executive orders, signalling acceleration over precaution. Strategic pillars include streamlining data centre permitting, removing perceived environmental constraints, and spreading the “American AI Technology Stack” to allies while advocating for American values in international standard-setting bodies ¹.

The Chinese Model: AI as Sovereign Infrastructure

State-Centric Strategy: China treats AI as sovereign infrastructure controlled for strategic goals under its “New Generation AI Development Plan” (2017), with ambitions to become the world’s primary AI innovation centre by 2030. Deep military-civil fusion ensures commercial advances systematically support defence modernisation ². Implementation occurs through state-backed investment funds, subsidised computing, national AI pilot zones, and the Digital Silk Road initiative exporting China’s technology stack and governance model globally ³.

The European “Third Way”: Regulation as Power

Rights-Based Framework: The EU AI Act establishes the world’s first comprehensive, legally binding AI regulation with risk-based classifications – prohibiting social scoring and manipulative AI systems while imposing strict obligations for high-risk applications in critical infrastructure, medical devices, recruitment, and law enforcement. The “Brussels Effect” leverages market size to export this rights-based framework as a de facto global standard, though critics worry stringent regulations may stifle innovation ⁴.

UK’s Strategic Pivot: Pro-Innovation Middle Path

Post-Brexit Positioning: The UK navigates between heavily regulated EU and largely deregulated US models, delegating governance to existing regulators (FCA, ICO, CMA) rather than creating overarching AI law. Five cross-sectoral principles guide safety, transparency, fairness, accountability, and redress ⁵. However, heavy reliance on American tech investment risks entrenching Big Tech power and compromising regulatory sovereignty, potentially reducing the UK to junior partner status despite AI superpower ambitions ⁶.

Control, Containment, and Information Integrity

Nations are deploying hard and soft power tools to gain strategic advantage, from semiconductor chokepoints to information sphere battles.

The Semiconductor Chokepoint

Export Control Strategy: The US weaponises its position in the global chip supply chain through strict controls on logic chips (16/14nm FinFET), DRAM, and manufacturing equipment exports to China, while prohibiting US persons from supporting advanced Chinese chip facilities ⁷. This creates a strategic paradox: short-term “compute bottlenecks” for Chinese AI firms may catalyse long-term push for technological self-sufficiency. The controversial 2025 arrangement requiring Nvidia/AMD to share 15% of China sales revenue with the US government marks a new transactional phase ⁸.

The Open-Source Dilemma

Innovation vs. Proliferation Tension: Open-source models accelerate research and provide competitive checks on tech giants but enable adaptation by malicious actors for disinformation, cyberweapons, and biosecurity threats. China’s AI ecosystem significantly relies on Western open-source models like Meta’s Llama, while Chinese firms produce world-class open-source models (Alibaba’s Qwen, 01.AI’s Yi) for global influence. Both UK and China support open-source principles despite underlying competition ⁹.

Information Warfare and Deepfakes

Synthetic Media Threat: AI tools produce highly realistic deepfake content at unprecedented scale, threatening electoral processes and public trust. China implements comprehensive bans on AI-generated fake news with mandatory labelling, while Western approaches target specific high-harm cases (non-consensual pornography, political manipulation) while protecting free speech ¹⁰. Technical solutions include digital watermarking (C2PA standard), content verification (Google’s SynthID), and “pre-bunking” educational initiatives ¹¹.

Energy Geopolitics and Resource Security

The exponential growth of AI introduces new hard-edged dimensions to geopolitics through competition for physical resources required to power it.

The Energy Imperative

Unprecedented Demand: Global data centre electricity demand will exceed 945 TWh by 2030 – greater than Japan’s annual consumption. US data centres will account for nearly half of new electricity demand growth, with connection delays exceeding two years forcing reactivation of retired coal plants. Single AI queries consume half a litre of water to be circulated for cooling, with tech companies reporting 20-34% usage increases, threatening climate commitments and raising electricity bills for consumers effectively subsidising Big Tech ¹².

National Resource Strategies

The US pursues “Build, Baby, Build!” – treating energy infrastructure expansion as security imperative while removing Clean Water Act and Clean Air Act constraints to prioritise unconstrained AI capacity. China’s “East Data, West Computing” strategy builds data centre clusters in western provinces with abundant renewable energy, leveraging cooler climates while promoting regional development ¹³. Tech companies invest billions in Small Modular Reactors (SMRs), with Microsoft, Google, Amazon, and Meta leading nuclear energy initiatives for reliable 24/7 carbon-free power ¹⁴.

Green AI as Competitive Arena

Europe makes environmental sustainability a core AI strategy pillar through mandatory energy/water consumption reporting and sustainability labels. The UK’s £10+ billion AI for Climate programme uses AI to improve energy efficiency and accelerate renewables transition, raising questions whether efficiency focus creates long-term advantage or regulatory drag ¹⁵.

Divergent Regulatory Frameworks

The global battle to set AI standards reflects fundamental differences in political values, economic priorities, and geopolitical ambitions.

EU AI Act: Rights-Based Global Standard

The Act establishes a four-tier risk classification with phased implementation: prohibitions became effective in February 2025, general-purpose AI rules applied from August 2025 and full compliance by 2026-2027. Its extraterritorial reach applies to any AI system in the EU market regardless of provider location ¹⁶.

US Patchwork Approach

Federal policy relies on voluntary NIST AI Risk Management Framework rather than comprehensive legislation, with Biden’s “AI Bill of Rights” superseded by Trump’s deregulatory “AI Action Plan.” Over 500 AI-related bills in state legislatures create compliance complexity ¹⁷.

UK Sector-Specific Model

Existing authorities interpret five cross-sectoral principles in an agile framework intended for context-aware, sector-specific rules. Parliamentary push for formal AI Authority through Private Member’s Bill challenges government approach ¹⁸.

China’s State-Controlled Framework

China implements nationwide rules for recommendation algorithms and generative AI, fostering innovation while maintaining information control. Requirements that AI content “adheres to core values of socialism” reflect dual objectives ¹⁹.

National Priorities and Public Trust

Long-term AI leadership requires strategic investment and public support, with trust emerging as critical enabler or inhibitor of deployment scale and pace.

Investment Strategies

The US democratises computing access through the National AI Research Resource (NAIRR) with $700+ million annual NSF funding focused on national security applications. China’s $47 billion “Big Fund” for semiconductor independence combines with AI Pilot Zones and subsidised compute vouchers through coordinated top-down direction ²⁰. The UK plans 20-fold expansion of AI Research Resource by 2030, establishes AI Growth Zones, and creates a National Data Library, though success depends heavily on foreign investment ²¹.

The Trust Divide

High-trust societies like China (72-83% public trust) and developing nations view AI as economic growth tool, enabling rapid deployment generating vast training data. Low-trust democracies show US (32-39% trust) and Western scepticism about job displacement, misinformation, and privacy erosion, creating natural deployment brakes through public consultations and regulatory oversight ²².

Public Engagement Strategies

The US shifts from Biden’s rights-focused “AI Bill of Rights” to Trump administration’s workforce development emphasis, with philanthropic funding supporting public interest AI. China promotes state-directed techno-optimistic narrative of AI as national rejuvenation driver through controlled “Responsible AI” presentations ²³. The UK’s National AI Strategy explicitly identifies public trust as core pillar, building “social license” through consultation reflecting societal values ²⁴.

Recent Developments and Strategic Outlook (2025)

The AI political landscape is evolving rapidly, with major implementations, escalating competition, and emerging cooperation efforts shaping the strategic environment.

Policy Implementation

EU AI Act enforcement began with February 2025 prohibitions on unacceptable-risk systems, August 2025 general-purpose AI rules and testing the “Brussels Effect” globally ²⁵. The UK has implemented its 50-Point Action Plan establishing AI Growth Zones and National Data Library while funding regulator AI capabilities. US-China competition intensifies through export control evolution, controversial chip revenue-sharing, and China’s notable progress in efficient architectures despite restrictions ²⁶.

Future Trajectories

The central US-China rivalry sees China closing performance gaps while excelling at broad-based adoption. EU tests whether regulatory influence pulls nations into rights-based orbit, while Global South alignment determines future geopolitical balance ²⁷. A critical five-year window will determine whether AI order becomes bipolar (US-China), tripolar (+ EU), or fragmented alliance landscape, with regulatory convergence competing against deregulatory models and fragile international cooperation channels building on Bletchley Declaration foundations ²⁸.

References: