On the Issues
Artificial intelligence is rapidly reshaping how we work, communicate, and protect our national security. While AI has the potential to deliver real benefits, it also poses serious risks if deployed without accountability, transparency, or public oversight. Public First Action advocates for AI policies that put people first—protecting families and workers, preserving democratic accountability, and ensuring the United States remains secure and competitive in a changing world.
Our Approach to AI Policy
Although artificial intelligence has the potential to improve lives in many ways, it poses real risks to jobs, families, national security, and public safety. Effective AI governance requires proactive safety standards, transparency and accountability mechanisms, and protections for whistleblowers alerting the public and policymakers to dangerous activities.
We must ensure that risks are identified and addressed before harm occurs, while holding AI developers accountable for the real-world impacts of their systems.
To identify and mitigate emerging risks, we need transparency from AI companies, and companies developing advanced AI should be required to inform policymakers and other authorities of what they are doing, before they do it.
Companies developing potentially harmful or dangerous AI should be required to evaluate the safety of those systems before they are deployed.
Whistleblowers who expose dangerous activity by AI companies should be legally protected from retaliation.
Accountability and Transparency
To maintain our edge in AI development and to mitigate national security risks, the federal government should restrict the export of advanced chips to countries of concern like China.
Stricter export controls should include restrictions on selling our adversaries more powerful chips than what they can produce domestically, location verification mechanisms on exported advanced chips, increased funding for anti-smuggling enforcement, and significant penalties for chip diversion.
Protect National Security
States are often best positioned to respond quickly to emerging harms. Federal standards shouldn't preempt stronger state protections.
Competition between states drives innovation in finding effective policies to protect the public from AI harms.
Preserve States’ Authority
Don’t Block State AI Laws
Congress has rejected proposals that would block states from regulating artificial intelligence. Preempting state AI laws without a strong federal standard would mean that states can’t protect families from harmful AI products.
AI already poses significant threats to economic stability, privacy, children, national security, and public safety.
There have been numerous cases of psychosis and teen suicide linked to chatbot use.
AI chatbots have allowed users to create child sexual abuse material.
AI can be used to create images and videos indistinguishable from reality.
These, and even more severe, risks and dangers are only growing as AI becomes more capable and widely deployed.
States have stepped in where Congress has not by passing laws to protect families from harms from AI products and to reduce the risks of future AI dangers.
States have passed laws that focus on transparency for the big AI companies and to protect children and other vulnerable populations.
Big Tech does not want to be subject to guardrails, so they have tried to get Congress to pass laws blocking states from regulating AI, replaced by either weak federal standards or none at all.
States serve as laboratories of democracy, driving innovation and competition to find the best AI policies, and blocking state AI laws undermines federalism and overrides states' rights.
Given the stakes involved, state AI laws should be preempted only if Congress passes strong nationwide rules covering those specific areas of AI policy, meeting or exceeding the state standards.
Accountability and Transparency
73% of Americans think AI companies should be liable for harms caused by their technology.
More than twice as many Americans worry the government will not go far enough in creating guardrails for AI than are concerned that the government will go too far.
Tragic stories about AI psychosis, self-harm, and children's mental health have raised the public's awareness about the need for AI safeguards.
Voters Want Safeguards, Not Preemption
Maintaining AI Export Controls
Artificial intelligence has the potential to reshape the global balance of power. The nation that leads in AI will have decisive advantages in economic competitiveness, military capability, and technological innovation. America leads for now, but maintaining that lead depends on our control of the powerful computer chips that make advanced AI possible.
AI leadership will determine who gains economic, military, and technological advantages, and whether those advantages prove durable or destabilizing.
The AI that powers commercial applications can also enable military targeting, autonomous weapons, global surveillance, and cyber attacks.
When the gap between competitors is narrow, there is greater pressure to rush deployment before systems are fully tested or controlled.
Whoever is ahead will have the power to shape global standards for AI in defense, intelligence, and critical infrastructure.
The AI Competition with China Matters
Advanced chips are the foundation of AI: Training powerful AI requires massive computing power from specialized processors.
The most advanced chips cost tens of thousands of dollars each, and it takes thousands of them to train and run cutting-edge AI systems.
The U.S. has a substantial advantage in advanced chips over China, giving us a lead in AI development.
Our Lead Depends on the Most Advanced Chips
China does not have the domestic capability to manufacture the most advanced chips and is unlikely to catch up in the foreseeable future.
By restricting chip sales to China, we force them to use slower, less capable processors and limit their ability to develop advanced AI for military and surveillance purposes.
There is broad, bipartisan support for maintaining and strengthening controls that prevent the most powerful chips and chip-manufacturing equipment from being exported to China.