Global AI Governance Fractures as Nations, States, and Markets Diverge on Safety and Growth
The global AI governance consensus is fracturing along three fault lines: between nations, between federal and state authorities, and between safety advocates and growth maximalists. The New Delhi AI summit produced an 88-nation declaration stripped of binding safety commitments, completing a three-year retreat from the enforceable guardrails envisioned at the 2023 UK summit. India's successful reframing around "democratization" signals that the Global South will increasingly set AI's normative agenda.
Singapore is emerging as a case study in how a small, open economy weaponizes AI strategy to offset geopolitical shocks. Budget 2026 commits S$37 billion to R&D, establishes a PM-chaired National AI Council, and launches a national space agency — all while bracing for Trump's new 15% tariff imposed under a never-before-used Cold War-era trade provision. The city-state's simultaneous investment in AI, space, and growth capital infrastructure represents a bet that technological sovereignty can buffer trade volatility.
In the United States, a constitutional collision is unfolding between federal deregulation and state-level AI accountability. The Trump administration threatens lawsuits and withholds broadband grants to block state AI laws, while Colorado, California, and Utah press ahead with safety and transparency requirements. This fragmentation creates a patchwork compliance landscape with no resolution in sight.
On a civilizational timescale, the signals are stark: AI is already restructuring creative labor (UNESCO projects 24% music revenue losses by 2028), rewriting geopolitical alliances, and challenging the Westphalian assumption that nation-states can govern transformative technologies within their borders. The question for this decade is whether humanity builds governance institutions that match AI's speed — or whether the absence of binding global rules becomes the permanent status quo.
1 · Do
1.1
Today · with Trade policy advisors, Supply chain leaders, Finance teams
Monitor the 150-day countdown on Trump's Section 122 tariff authority and assess exposure across all supply chains with US trade dependencies, modeling scenarios for extension, congressional legislation, or expiration to prepare adaptive responses. (Addresses: trade resilience and geopolitical risk)
This Week · with AI ethics leads, Legal counsel, International policy teams
Review the New Delhi AI declaration's voluntary framework and assess how the absence of binding global safety commitments affects organizational AI deployment risk, particularly regarding open-source model adoption and cross-border data governance in markets where no enforceable standards exist. (Addresses: AI governance and responsible deployment)
This Month · with Legal and compliance teams, Government affairs, Product teams deploying AI in the US
Assess the emerging US state AI regulatory patchwork — Colorado's AI Act (effective summer 2026), California's Digital Dignity Act, and Utah's frontier model safety bill — and develop a compliance matrix that accounts for both state enforcement and potential federal preemption lawsuits. (Addresses: regulatory compliance)
This Quarter · with Strategy teams, Regional business development, R&D leadership
Engage with Singapore's National AI Council framework and RIE2030 funding opportunities to evaluate whether the city-state's AI mission areas — advanced manufacturing, connectivity, finance, healthcare — align with organizational expansion or partnership strategies in Southeast Asia. (Addresses: strategic growth and market positioning)
This Quarter · with HR leadership, Creative and content teams, Strategy planning
Prepare an internal assessment of generative AI's impact on creative and knowledge-work functions, using UNESCO's 24% music revenue decline projection as a baseline, to identify roles and revenue streams at risk of displacement and develop workforce transition plans. (Addresses: workforce strategy and long-term planning)
Global AI safety consensus collapses as New Delhi declaration drops binding commitments The fourth global AI summit produced a declaration signed by 88 nations with zero enforceable safety measures — a complete reversal from the 2023 UK summit's focus on existential risk. India's successful 'democratization' framing, endorsed by UN Secretary-General Guterres, signals that AI governance is shifting from Western safety-first norms to Global South adoption-first priorities. Over a 10-year horizon, this vacuum makes voluntary industry self-regulation the de facto global standard, with profound implications for whether frontier AI development faces any meaningful external checks.
2.2
Trump's tariff escalation under Section 122 tests constitutional limits of executive trade power After the Supreme Court blocked IEEPA-based tariffs, Trump pivoted to Section 122 of the 1974 Trade Act, imposing a 15% global levy with a 150-day limit and no congressional approval. Singapore's DPM Gan called exemption negotiations 'very difficult' with blanket tariffs. This constitutional cat-and-mouse between executive power and judicial review creates a structurally unstable trade environment that forces every export-dependent economy to build resilience against perpetual policy oscillation.
2.3
Singapore commits record S$37 billion to research as AI becomes national strategic pillar The RIE2030 plan represents a 32% increase over RIE2025, with AI missions in manufacturing, connectivity, finance, and healthcare overseen by a PM-chaired National AI Council. OCBC's chief economist called it 'an all-out coordinated and multifaceted push.' Singapore's model — deploying rather than developing frontier AI — offers a template for mid-sized nations seeking technological relevance without competing directly with US-China compute dominance.
2.4
US federal-state AI governance clash creates patchwork regulatory landscape with global spillovers The Trump administration threatens DOJ lawsuits and $42.5 billion in broadband grant withholding against states like Colorado, California, and Utah that regulate AI. Colorado's AI Act takes effect summer 2026; California introduced the Digital Dignity Act and multiple children's safety bills. The ACLU calls the federal approach 'a hodgepodge of faulty legal theories.' For any organization operating across US states, compliance planning now requires scenario modeling for both federal preemption and continued state fragmentation.
2.5
UNESCO quantifies AI's economic disruption to creative industries across 120 countries Music creators face projected 24% revenue losses and audiovisual workers 21% by 2028. A 'creative digital divide' compounds the damage: 67% digital skills in developed nations versus 28% in developing ones. These figures represent the first large-scale quantification of AI's labor displacement in the creative economy — a leading indicator for how generative AI will restructure employment across knowledge-work sectors over the next decade.
2.6
Singapore launches space agency to position itself in $1.8 trillion orbital economy NSAS officially launches April 1, 2026, building on SGD $200 million in space R&D and an ecosystem of 70 companies and 2,000 professionals. The global space economy is projected to reach $1.8 trillion by 2035. Singapore's bet is that space is the next infrastructure layer — like undersea cables or semiconductor fabs — where a small nation with regulatory clarity and deep-tech expertise can command outsized influence.
2.7
AI-powered brain-computer interfaces and mental health tools reach deployment stage Cognixion's noninvasive brain-computer interface enables mind-controlled apps for nonverbal individuals; Dartmouth's Evergreen uses biometric data for personalized mental health support built by 100+ students and faculty. These are no longer lab demonstrations but field-deployable systems. They represent AI's capacity to expand human agency — the constructive counterweight to the displacement narrative dominating policy discussions.
2.8
States deploy AI for compliance enforcement while simultaneously regulating AI itself Montana uses AI dashboards to review campaign finance filings; Hawaii deploys AI for investigations. Yet AI-generated citizen complaints frequently reference incorrect state requirements, creating enforcement noise. This recursive dynamic — regulators using AI to regulate AI — previews governance challenges that will intensify as AI becomes embedded in institutional decision-making at every level.
3 · What changed
3.1 · Regulatory Changes
New Delhi AI Summit Yields 88-Nation Declaration Without Binding Safety Commitments
Some 88 countries including the US, UK, EU, and Russia signed an India-led AI declaration at the fourth global AI summit in New Delhi on February 21, 2026. The declaration recognized 'the importance of security in AI systems' but included no binding safety measures, relying instead on voluntary and industry-led approaches. Indian PM Narendra Modi successfully reframed the summit around 'democratization' and 'wide-scale adoption,' shifting away from the safety focus of the 2023 UK summit. The declaration boosted open-source AI, stating that open-source approaches 'can contribute to scalability, replicability, and adaptability.' UN Secretary-General Antonio Guterres said the future of AI should not be decided by 'a few billionaires.' Neither the US nor Chinese heads of state attended.
3.2 · Regulatory Changes
Trump Imposes 15% Global Tariff Under Section 122 After Supreme Court Rebuke
US President Donald Trump raised the global tariff on US imports to 15% on February 21, 2026, using Section 122 of the Trade Act of 1974, which allows tariffs of up to 15% for 150 days without congressional approval. The move followed a US Supreme Court ruling that Trump lacked authority under the International Emergency Economic Powers Act (IEEPA) to impose his previous tariff program. Singapore's Deputy Prime Minister Gan Kim Yong described the situation as 'very fluid' and a 'stark reminder' of the unpredictable operating environment. Singapore will seek clarity from US counterparts on implementation and tariff refund processes. The US ran a goods trade surplus of US$3.6 billion with Singapore in 2025. DPM Gan warned it would be 'very difficult' to negotiate exemptions if tariffs apply to all countries.
3.3 · Technology Disruption
Singapore Budget 2026 Positions AI as Central Economic Strategy with S$37 Billion R&D Push
Singapore's Budget 2026, delivered by PM Lawrence Wong on February 12, places AI at the center of the nation's economic strategy with a record S$37 billion commitment under the RIE2030 plan — a 32% increase over RIE2025. PM Wong will chair a new National AI Council, and the government announced AI missions targeting four sectors: advanced manufacturing, connectivity, finance, and healthcare. A dedicated AI park at One-north and a Champions of AI programme are also planned. OCBC's chief economist called it 'an all-out coordinated and multifaceted push to make AI central to Singapore's economic and workforce strategy for the coming years, if not decades.' GDP growth is projected at 2-4%, moderating from 4.8% in 2025. A Growth Capital Workgroup chaired by Minister Chee Hong Tat will recommend measures to strengthen Singapore as a growth capital hub, including S$1 billion for the Startup SG Equity scheme.
3.4 · Regulatory Changes
Trump Administration Threatens Lawsuits Against States Over AI Regulation
The Trump administration plans to sue states with AI laws it deems 'burdensome,' following a December executive order directing the Department of Justice to act on interstate commerce grounds. The White House sent a memo opposing Utah's HB 286, which requires frontier AI developers to publish safety and child protection plans, calling it 'an unfixable bill.' Colorado's AI Act, requiring 'reasonable care' against discrimination for high-risk AI systems, is scheduled to take effect in summer 2026. California introduced the Digital Dignity Act (SB 1142) and multiple placeholder bills on AI chatbots and children's safety. States could face lawsuits and withholding of $42.5 billion in BEAD broadband grants. The ACLU called the administration's plan 'a hodgepodge of faulty legal theories.' Utah's Republican sponsor and Governor Cox have defended states' rights to regulate AI independently.
3.5 · Market Trends
UNESCO Reports AI Will Cut Music Creator Revenues 24% by 2028
UNESCO's global report covering 120+ countries projects generative AI will drive income losses of 24% for music creators and 21% for audiovisual workers by 2028. The report, released February 18, identifies a 'creative digital divide' where 67% of people in developed countries possess essential digital skills versus just 28% in developing nations. UNESCO Director-General Khaled El-Enany outlined more than 8,100 policy measures and called for urgent coordinated action to protect creators' rights and strengthen regulatory frameworks. The agency warned that streaming platform dominance and opaque algorithms are widening disparities among creators, particularly in the Global South.
3.6 · Geographic Expansion
Singapore Launches National Space Agency to Capture $1.8 Trillion Space Economy
Singapore announced the creation of the National Space Agency of Singapore (NSAS) on February 2, 2026, with an official launch date of April 1 under the Ministry of Trade and Industry. Led by veteran public servant Ngiam Le Na, the agency builds on the Office for Space Technology and Industry (OSTIn). Singapore has committed over SGD $200 million to space R&D since 2022, targeting satellite remote sensing, space sustainability, and microgravity research. The country already hosts approximately 70 space companies employing around 2,000 professionals. The global space economy is projected to reach $1.8 trillion by 2035.
3.7 · Product Innovation
AI Applications Advance Brain-Computer Interfaces and Mental Health Support
TechNet highlighted several AI applications addressing human challenges. Cognixion develops noninvasive brain-computer interfaces for nonverbal individuals with disabilities, enabling mind, eye, and head-controlled wireless assisted reality applications. Dartmouth's Evergreen platform, built by over 100 students and faculty, provides AI-powered personalized mental health support using biometric data, routines, and academic schedules. Google's AI tools — DeepPolisher, DeepVariant, and DeepConsensus — support the Vertebrate Genomes Project, sequencing genetic data for 13 endangered species. Singapore-based MangaChat, using IP from Tohoku University and Panasonic, builds AI-powered journaling for individuals with ADHD and autism.
4 · What we could and could not see
0 topic searches. 24 articles from 22 outlets, 2 of them from sandpipercomms.com. 4 are not in a story.
3 of 7 stories rest on a single outlet. Treat those as reported, not confirmed.
