As submitted to the Office of Science and Technology Policy in response to a Request For Information asking for input for their upcoming US AI Action Plan

Introduction

Artificial Intelligence (AI) is more than a technological breakthrough—it is a transformative force shaping the future economy, security landscape, global power dynamics, and daily life. The US, along with its international allies and private sector leaders, must navigate a pivotal moment where their choices will determine whether AI advances democracy and human progress or fuels authoritarianism and societal instability. To ensure AI remains a force for good, a comprehensive strategy must prioritize:

1. Keeping AI and AI-enabled technologies under human control.
2. Securing US AI leadership and economic competitiveness.
3. Retaining collaborative government oversight of AI Industry.
4. Addressing societal risks posed by AI while strengthening democratic institutions.
5. Expanding the benefits of AI globally.

These challenges require decisive action to prevent AI from being exploited for dominance and control. With strategic leadership, the US and its allies can harness AI to uphold democracy, drive economic growth, and promote global stability.

1. Keeping AI and AI-Enabled Technologies Under Human Control

The rapid evolution of AI, particularly the potential emergence of Artificial General Intelligence (AGI), presents both opportunities and challenges that require strategic, proactive oversight to ensure AI systems remain aligned with human preferences, dignity, security, and economic growth. While the possibility and timeline for AGI remain widely debated, the need for AI systems to operate under responsible human oversight is clear. Rather than imposing excessive regulatory constraints, the responsible development of AI requires a collaborative approach among policymakers, industry leaders, and global partners to ensure human oversight, ethical use, and continued innovation.

Moreover, the principles of responsible AI oversight must also be extended to AI-enabled technologies such as biotechnology, nanotechnology, and other scientific advancements that can be leveraged for economic competitiveness, while ensuring safeguards against misuse. Just as collaborative AI governance is essential to advancing US leadership and security, parallel safeguards must support the responsible integration of AI across other technology sectors.

In addition to maintaining “human-over-the-loop” (where humans provide necessary input or oversight), especially with national security and research AI models, it is important to keep a “human-in-the-loop” for high-risk AI applications, where direct human decision-making is essential to prevent unintended harm (e.g., autonomous weapons, medical AI) and “human-on-the-loop” where AI can be allowed to operate mostly autonomously, but humans must be able to intervene when necessary (e.g., self-driving cars, AI-assisted financial systems). Advances in "Human-guided AI" where AI can function with minimal supervision in low-risk environments can maximize efficiency and productivity.

Action Steps to Ensure Human Control Over AI and AI-Enabled Technologies

1. Strengthen the Role of the US AI Safety Institute (AISI) – The AISI, a unit of US Commerce Department’s National Institute of Standards & Technology (NIST) should be the lead entity in conducting independent risk assessments to evaluate high-risk AI models. It should work closely with private industry, ensuring that AI safety efforts align with human control, economic growth and national security imperatives.
2. Ensure Mandatory Shutdown Mechanisms – AI systems must be designed with fail-safe mechanisms that allow immediate deactivation if they deviate from intended functions or pose security risks. This is particularly critical for autonomous weapons, financial trading AI, AI-powered infrastructure control, and medical AI diagnostics.
3. Government Authority to Temporarily Suspend Unsafe AI Systems – While industry-led safeguards are crucial, the government must retain authority to temporarily suspend AI systems that pose serious safety or security risks, like National Highway Traffic Safety Administration vehicle recalls. This last-resort measure ensures public safety when industry mechanisms fail. This temporary suspension mechanism would require necessary corrections before AI systems can be redeployed in high-risk environments.
4. Define Global AI Safety Standards – The US should leverage its AI leadership to shape global AI safety frameworks rather than adopting overly restrictive compliance models. Through initiatives like the Artificial Intelligence Safety Institute Consortium (AISIC) and the International Network of AI Safety Institutes (INASI), the US can engage allied nations, researchers, and industry leaders to establish voluntary, adaptable AI safety standards that align with US interests while balancing security and economic priorities.
5. Promote AI Alignment Research – The US should invest in research on AI interpretability and value alignment to ensure AI systems operate within security constraints. Initiatives such as Robust Open Online Safety Tools (ROOST) that promote open-source safety software and tools should be encouraged. 
6. Avoid Miscalculations with Weapons. Given the increasing role of AI in defense, the US should explore strategic stability agreements with China and Russia to ensure that AI is not misused for nuclear command and control operations. Such agreements should not limit US AI defense advancements but can help prevent misunderstandings or escalatory risks in military AI applications. The US should engage in dialogue with other leading AI powers to promote stability and prevent miscalculations in military AI applications.
7. Modernize Department of Defense (DoD) Policies – As warfare automation accelerates, "human in the loop" will increasingly mean embedding human judgment in decision-making rather than direct intervention in every decision cycle. To ensure AI systems remain controlled and non-escalatory, human oversight must be hardwired into system design before conflicts arise. The DoD should implement rigorous modeling, simulation tools, and real-time AI decision support systems, ensuring that US defense remains superior while AI systems function reliably in mission-critical applications and that AI decision support avoids escalation bias.

Rather than approaching AI regulation with heavy-handed government intervention, the US must lead through a balanced strategy that strengthens AI governance while maintaining flexibility for innovation. By shaping global AI safety standards, supporting public-private risk management frameworks, and modernizing AI oversight in national security applications, the US can mitigate risks without compromising its leadership in AI-driven economic and military capabilities.

2. Securing US AI Leadership and Economic Competitiveness

AI is not just a technological race—it is a contest of governance models that will shape global power, economic growth, and individual freedoms. The US must ensure that democratic AI that empowers citizens remains stronger, safer, and more influential than its authoritarian counterparts that seek to control citizens. Only by staying ahead in AI can the US sustain it economic competitiveness that underwrites its prosperity, security and values. 

Action Steps to Strengthen US AI Leadership and Economic Competitiveness

• Offense

1. Invest in AI Talent Pipeline – China graduates twice the number of Science, Technology, Engineering, and Math (STEM) PhDs as the US. With the US ranking near the bottom in the percentage of graduates pursuing STEM fields, it has long relied on attracting the world’s best and brightest. Visas for Chinese students and academics have fallen by two-thirds since peaking in 2015. A recent surge in students from India is taking up some of the slack. It is important that policies and a welcoming environment continue to attract the global AI talent essential to US tech leadership while also investing more in STEM education for American students.
2. Provide Access to High-Performance Compute Power – Establish a national AI compute network to support research universities. Increase domestic semiconductor production and high-performance computing (HPC) resources to support AI research and private-sector innovation. Gain a decisive lead in emerging technologies that promise faster and greener compute power, such as optical computing (using laser light waves), neuromorphic computing (modeled after systems in the human brain), and quantum computing (using subatomic particles).
3. Secure Data – Secure access to unbiased, high-quality data for US AI researchers. Consider which datasets should be restricted from strategic competitors. Counter China's growing control over the data flowing through global telecommunications (Huawei, ZTE) and logistics (LOGINK) software, while promoting secure, US-aligned alternatives.
4. Fund Research – Fully fund research as authorized in the CHIPS and Science Act, prioritizing AI and other technologies that will shape tomorrow like quantum, biotech and nanotechnology. According to the Organization for Economic Cooperation and Development (OECD), China recently overtook the US in research funding at universities and government entities. Unless funding trends are reversed China will soon overtake the US in total research funding. The US must ensure American institutions remain the world’s AI research hubs.
5. Invest in AI for Economic Competitiveness – Support AI-driven innovation across key industries (healthcare, manufacturing, energy, finance) by incentivizing AI adoption in small businesses and expanding AI-driven workforce training programs. The goal should be for AI to enhance productivity and job creation to ensure continued opportunities for American workers.

• Defense

6. Leverage AI for National Security Superiority – Expand AI-driven cybersecurity measures, real-time threat detection, and AI-powered intelligence capabilities. Modernize DoD procurement to rapidly integrate AI-driven military applications, ensuring US military dominance in AI-powered defense systems.
7. Secure AI Supply Chain – Reduce dependence on adversarial nations for rare earth materials, semiconductors, and cloud computing infrastructure. Mitigate risks posed by foreign adversaries exploiting digitally driven devices and information platforms to safeguard national security, economic stability, and democratic integrity. Establish secure, US-led AI supply chains to prevent exploitation by strategic competitors and strengthen US economic resilience.
8. Calibrate AI Export Controls – Restrict AI-enabling technology exports that could accelerate competitors’ military capabilities while ensuring export policies do not inadvertently accelerate their self-sufficiency, pairing restrictions with strategic tech-sharing among democratic allies and deployment to the greatest extent possible to emerging markets. AI export policies should balance security needs with economic competitiveness.
9. Enhance AI Model Security & Oversight – Both open-weight and closed AI models pose security risks. Open-weight models are vulnerable to adversarial fine-tuning and misuse, while closed models risk hidden biases and covert data collection. The AI Dispersion Framework has mitigated some threats by restricting AI compute power and frontier model access (though it overreaches potentially giving China the advantage in emerging markets). Develop safeguards to prevent misuse of open-weight AI and to restrict usage of foreign-developed AI in sensitive sectors as necessary to protect national security.
10. Influence International AI Standards – The US must lead in shaping global AI governance to ensure international standards protect individual liberties, foster innovation, and strengthen economic and national security. As China exports its state-controlled AI model and the EU imposes restrictive regulations, fragmented policies risk undermining American competitiveness and personal freedoms. By advancing transparent, market-driven AI standards, collaborating with allies and industry, and aligning AI export controls with strategic interests, the US can champion an AI future that empowers citizens and reinforces democratic values while maintaining technological leadership.

By prioritizing AI investment, talent development, and infrastructure security, the US can maintain global AI leadership without unnecessary regulatory burdens. A pro-innovation approach will ensure that AI remains an engine of American economic strength and military superiority. With a strong focus on both economic expansion and national security, the US will continue setting the global AI agenda—rather than reacting to authoritarian models.

3. Retaining Collaborative Government Oversight of AI Industry

As AI transforms industries and global competition intensifies, America’s tech leaders must remain strategic assets in national security and economic strength. Unlike China’s centralized control over firms like Alibaba, Tencent, and Huawei, the US has succeeded by fostering innovation through a free-market approach. Ensuring AI advances national security and economic priorities requires a balanced, pro-innovation strategy that encourages responsible corporate leadership without stifling growth. Instead of burdensome government intervention, AI oversight should be a collaborative effort between industry, government, and research institutions to strengthen US AI competitiveness while preserving national interests.

Action Steps for AI Corporate Oversight

1. Ensure AI Innovation Aligns with National Security – Work with AI firms to ensure US leadership in AI-driven defense, cybersecurity, and critical infrastructure while preventing partnerships that compromise US strategic interests.
2. Foster Competitive AI Markets – Avoid unnecessary regulatory overreach while ensuring a dynamic and competitive AI sector that rewards innovation and market-driven expansion.
3. Maintain AI Infrastructure Sovereignty – Strengthen domestic AI infrastructure through public-private partnerships that expand compute access and data set access to research universities and entrepreneurs. Ensure America—not foreign adversaries—dominates AI supply chains.
4. Encourage Responsible AI Governance by Tech Firms – Work with AI firms to develop voluntary AI safety frameworks, ensuring responsible innovation without heavy-handed government mandates. Support free speech principles while countering AI-driven disinformation.

The US must resist overreach that could stifle AI progress and instead leverage industry strengths to keep AI development aligned with national priorities. By fostering a strong public-private partnership, America can preserve innovation, economic leadership, and global AI ascendency—without adopting China’s heavy-handed control over its tech sector.

4. Address Societal Risks Posed by AI and Strengthen Democratic Institutions

AI has the potential to revolutionize industries and enhance decision-making. However, it also introduces significant societal risks, including biases in decision-making, economic displacement, erosion of privacy, and the spread of disinformation. If not carefully managed, AI could exacerbate socioeconomic inequalities and undermine democratic institutions. Ensuring AI benefits society while avoiding unnecessary restrictions is critical to maintaining public trust, fostering economic prosperity, and countering authoritarian AI models.

According to the 2024 Edelman Trust Barometer, global trust in AI companies declined from 61% to 53% over the past five years, with US trust levels dropping more sharply from 50% to 35% during the same period.  America trusting AI less than the rest of the world could hinder its ability to maintain its leadership in AI innovation. To reverse this trend and ensure AI's beneficial integration, it is vital to address societal risks while enhancing public understanding of AI's role in national competitiveness and the risks of ceding leadership to nations like China. As overly restrictive regulations could hinder progress, potentially allowing authoritarian regimes to gain an edge in AI development, a balanced approach is required.​

Action Steps to Address AI’s Societal Risks

1. Ensure AI Fairness & Prevent Discrimination – Support industry-led initiatives to reduce AI bias and enhance system reliability, avoiding costly and restrictive government audits.
2. Mitigate AI’s Impact on Employment – Expand AI workforce training and job transition programs, ensuring that workers benefit from AI-driven productivity rather than being displaced by it. While OECD countries spend 4.9% of GDP on education, only a small fraction of one percent is spent on adult education. Incentivize private-sector and academic investment in AI reskilling rather than relying solely on government programs.
3. Strengthen AI Transparent Government Use – Promote AI usage to provide 24/7 government services in a transparent manner, enhancing the availability and ease of accessing government services while maintaining trust and accountability.
4. Enhance Privacy & Data Protection – Encourage industry-driven privacy safeguards that protect consumers while allowing AI-driven services to thrive.
5. Combat AI-Powered Disinformation While Preserving Free Speech – Work with AI firms to advance voluntary content labeling, misinformation detection, and deepfake protections—without empowering censorship frameworks that could restrict political discourse.
6. Expand AI Literacy & Public Awareness – Support AI education initiatives that help the public understand and responsibly use AI, recognizing its role in job creation, national security, and global competitiveness. Encourage programs that equip individuals with the skills to navigate AI-driven technologies while fostering critical thinking about AI’s opportunities and risks.
7. Fair Value for Creators: Support market-driven agreements ensuring fair compensation for journalists, photographers, and other creators whose work trains AI models, balancing innovation with intellectual property rights while avoiding restrictive regulations.
8. Impact on Children: Work with AI firms to develop voluntary safeguards protecting children from harmful effects as AI shapes how they learn, play, and interact.
9. Environmental Sustainability: Encourage industry-driven advancements in AI energy efficiency, ensuring that AI progress aligns with sustainable development.
10. Mental and Social Health: Partner with tech companies to research AI’s influence on mental health and implement best practices.

Rather than overregulating AI and stifling growth, the US should strengthen public confidence in AI through private-sector leadership, voluntary compliance, and strategic public-private collaboration. By encouraging AI innovation, enhancing workforce readiness, and ensuring responsible AI governance, America can foster trust in AI while maintaining its global edge.

5. Extending AI’s Benefits to the World

AI is a strategic asset that enhances US economic influence, strengthens alliances, and counters authoritarian models of AI governance. The US must ensure that AI is not only a tool of economic and military power but also a driver of global prosperity, security, and democratic values. China’s Digital Silk Road is aggressively exporting AI-driven surveillance, censorship, and state control, embedding authoritarian AI models into developing economies. If left unchecked, China will dictate the future of AI governance, shaping global norms to favor state control over individual freedoms and economic independence. The US must act decisively to strengthen AI partnerships with allies, promote free-market, citizen-empowering AI ecosystems, and provide a compelling alternative to authoritarian AI exports.

Key Steps to Democratize AI Globally

1. Support AI Infrastructure – Encourage AI adoption in allied and emerging economies through public-private partnerships, investment incentives, and financing from the US International Development Finance Corporation (DFC) and the Export-Import Bank. By expanding US AI infrastructure abroad, America can counterbalance China’s Digital Silk Road and establish a free-market, empowering AI ecosystem.
2. Expand Global AI Education & Workforce Training – Position American universities and tech companies as the premier hubs for AI education and training. Strengthen US-led AI certification programs, university partnerships, and technical training initiatives in developing economies to solidify US leadership in AI knowledge transfer.
3. Leverage AI for Sustainable Development – Encourage US AI-driven innovations in healthcare, crisis response, and sustainability to drive economic growth in partner nations while reinforcing democratic norms. Expand AI-driven agriculture, logistics, and fintech solutions to strengthen global supply chain resilience and economic stability.
4. Provide Open and Democratic AI Solutions – Provide AI solutions that enhance security, economic growth, and digital freedom while preventing authoritarian AI models from embedding state surveillance into global infrastructure. Support democratic AI governance models that align with US economic and security interests.
5. Strengthen Cultural & Linguistic AI Applications: Encourage AI applications that promote linguistic diversity and cultural preservation, ensuring AI systems reflect the values of diverse global communities while maintaining US technological leadership in AI-driven communication tools.

By expanding AI infrastructure investment, strengthening AI education, and providing a superior alternative to China’s AI exports, the US can position itself as the global leader in AI development. This pro-innovation, pro-democracy AI strategy will strengthen alliances, create new economic opportunities, and reinforce the US as the preferred AI partner for developing nations. The US must act swiftly and decisively to ensure AI benefits all nations while securing American economic and geopolitical leadership in AI-driven technologies.

Conclusion

AI is the defining technology of the 21st century, and how it is developed, regulated, and deployed will shape the world for decades to come. The US must lead—not just to maintain technological superiority, but to lead in ensuring that AI serves humanity rather than becoming a tool for authoritarian expansion. By implementing safeguards to keep AI under human control, investing America’s AI leadership, ensuring collaborative government oversight, mitigating AI’s societal risks, and extending AI’s benefits worldwide, the US can secure a future in which AI is a force for prosperity, security, and freedom.

AI leadership is not just a technological race—it is a competition for the future of global influence, economic power and human progress. The US must take bold action to secure its position as the world’s AI leader, ensuring that AI strengthens free markets, safeguards national security, and fuels liberty and prosperity for generations to come. The time to act is now.