All Articles

The public deserves a say over the values of government-procured AIs.

public reasoning fidelity, government AI procurement, AI model character, ideological neutrality, Commission on Public AI Use, frontier AI models, model selection, democratic oversight of AI, US Sentencing Commission, Administrative Conference of the United States, deliberative polls, policy analysis sycophancy, AI governance, public AI values

One of the most valuable things governments can build today is the capacity to govern advanced AI competently in the future.

radical optionality, AI governance, transformative AI, frontier AI models, AI regulation, whistleblower protections, reporting requirements, transparency mandates, UK AI Security Institute, information-gathering authorities, state AI laws, model evaluations, tort liability, management-based regulation, private governance

American leaders agree that the AI race will shape the balance of power with China. But they can’t agree on how to ensure the technology advances American values.

Sino-American AI competition, AI race with China, American values, Chinese values, breakaway tech dominance, encoded values, emergent control regimes, democratic AI, techno-authoritarianism, AI governance, AI diplomacy, nuclear arms control, Global South, lethal autonomous weapons, President’s Council on AI

For decades, no nuclear power could disarm its rivals without provoking devastating retaliation. A large AI lead could change that.

AI capabilities gap, nuclear deterrence, mutual assured destruction, counterforce strike, nuclear primacy, launch on warning, missile defense, nuclear submarines, AI-assisted R&D, aerospace manufacturing, satellite constellations, strategic stability, arms control, military megaprojects, nuclear taboo

US restrictions on frontier AI would have come eventually, but few expected sudden export controls. They could be Europe's wake-up call on AI sovereignty.

Anthropic, Fable 5, Mythos 5, export controls, Europe AI sovereignty, frontier AI access, US restrictions, AI governance, compute capacity, European Union, AI Act, cybersecurity, middle powers, Trump administration, European dependence

AI’s economic impacts will unfold through several waves, with different policy approaches relevant to each phase.

Researchers and policymakers are fixated on the fear of AI launching nuclear weapons—to the neglect of more realistic threats.

AIs with access to all our data will soon be able to vouch for us to others. As people come to trust AI judgments of character, not sharing one will look suspicious.

Western AI safety treatises are surprisingly well-received in Chinese tech media. What does this mean for international AI policy?

AI forecasts span a range of potential futures, from economic stagnation to explosive growth. The divergence traces to three specific assumptions—each generating predictions we can already test.

Governments should set positive incentives for AI safety. Here are four approaches.

catalytic regulation, AI safety, positive incentives, deregulation, US AI policy, tax credits, procurement incentives, safety R&D, safety certification, organizational safety culture, high reliability organizations, AI governance, benchmarking and audits, prestige incentives

Current export rules focus on keeping chips a generation behind. They should focus on keeping America's total compute ahead.

export controls, Nvidia H200, AI processors, relative compute advantage, US-China tech competition, semiconductor exports, Commerce Department policy, chip performance thresholds, Blackwell generation GPUs, Huawei Ascend 910C, AI data-center compute, national security tradeoffs, chip export licensing, AI chip smuggling, relative advantage framework

Embodied AI is arriving faster than the regulations meant to govern it. If we start now, that’s a problem we can still fix.

embodied AI, humanoid robots, home robots, Figure 03, Unitree R1, robotics regulation, Consumer Product Safety Commission (CPSC), privacy laws CCPA/CPRA, Illinois BIPA, FTC COPPA, product liability for AI, NIST AI Risk Management Framework, ISO 13482, incident reporting for AI systems

AI is already taking jobs, but that is only one facet of its complex economic effects. Price dynamics and bottlenecks indicate that automation could be good news for workers — but only if it vastly outperforms them.

artificial intelligence, automation, job displacement, labor markets, productivity gains, price effects, bottlenecks, Baumol's cost disease, agricultural mechanization, computers and software, wage share, economic growth, full automation, inequality

Shaped by a different economic environment, China’s AI startups are optimizing for different customers than their US counterparts — and seeing faster industrial adoption.

China AI startups, US AI startups, AI investment gap, Hong Kong IPOs, Biren Technology, Zhipu AI, MiniMax, OpenAI Stargate, AI infrastructure spending, inference efficiency, Mixture-of-Experts models, industrial AI deployment, manufacturing AI adoption, enterprise AI solutions, AI monetization models

Modern memory architecture is vital for advanced AI systems. While the US leads in both production and innovation, significant gaps in export policy are helping China catch up.

high-bandwidth memory, HBM, DRAM, AI chips, GPU packaging, export controls, Bureau of Industry and Security, BIS, U.S.-China tech competition, semiconductor manufacturing equipment, FDPR, ASML immersion DUV lithography, SK Hynix, Samsung, Micron

Traditional insurance can’t handle the extreme risks of frontier AI. Catastrophe bonds can cover the gap and compel labs to adopt tougher safety standards.

frontier AI, extreme AI risk, catastrophic AI events, AI liability, liability insurance, catastrophe bonds, cat bonds, insurance-linked securities, capital markets, AI regulation, AI safety standards, third-party audits, catastrophic risk index, tail risk, systemic risk

The White House is betting that hardware sales will buy software loyalty — a strategy borrowed from 5G that misunderstands how AI actually works.

AI automation threatens to erode the “development ladder,” a foundational economic pathway that has lifted hundreds of millions out of poverty.

A growing body of evidence means it’s no longer tenable to dismiss the possibility that frontier AIs are conscious.

Community Notes offers a better model — where citizens, not corporations, decide what “aligned” means.

AI alignment, attentiveness, Community Notes, Taiwan, Audrey Tang, model specification, deliberative governance, epistemic security, portability and interoperability, market design, Polis, reinforcement learning from community feedback, social media moderation, civic technology

A new framework suggests we’re already halfway to AGI. The rest of the way will mostly require business-as-usual research and engineering.

AGI, artificial general intelligence, AGI definition, GPT-5, GPT-4, visual reasoning, world modeling, continual learning, long-term memory, hallucinations, SimpleQA, SPACE benchmark, IntPhys 2, ARC-AGI, working memory

By learning our views and engaging on our behalf, AI could make government more representative and responsive — but not if we allow it to erode our democratic instincts.

AI political proxy, direct democracy, generative social choice, ballot initiatives, voter participation, democratic representation, AI governance, Rewiring Democracy, Bruce Schneier, Nathan E. Sanders, policy automation, civic engagement, rights of nature, disenfranchised voters, algorithmic policymaking

China’s new AI safety body brings together leading experts — but faces obstacles to turning ambition into influence.

China AI Safety and Development Association, CnAISDA, China AI safety, World AI Conference, Shanghai AI Lab, Frontier AI risk, AI governance, international cooperation, Tsinghua University, CAICT, BAAI, Global AI Governance Action Plan, AI Seoul Summit commitments, Concordia AI, Entity List

A response to critiques of Mutually Assured AI Malfunction (MAIM).

AI deterrence, Mutually Assured AI Malfunction, MAIM, Superintelligence Strategy, ASI, intelligence recursion, nuclear MAD comparison, escalation ladders, verification and transparency, redlines, national security, sabotage of AI projects, deterrence framework, Dan Hendrycks, Adam Khoja