Policy
Practical public policies that preserve innovation while establishing responsible safeguards for increasingly powerful artificial intelligence.
01
America should continue building, researching and deploying advanced artificial intelligence. Falling behind is not a safety strategy.
Our goal is American technological leadership combined with world-leading safety standards. Build here. Innovate here. Protect people here.
02
Advanced AI can dramatically increase human capabilities. Those capabilities must not make it easier for terrorists, hostile governments or criminals to create catastrophic harm.
We support strong safeguards around the highest-risk applications of AI, including biological and chemical threats, advanced cyberattacks, critical infrastructure, autonomous weapons, nuclear command and control, and dangerous capabilities of frontier AI systems. Innovation can move quickly without ignoring national security.
03
Some decisions should never become unaccountable simply because an algorithm made them.
When AI affects life, liberty or fundamental rights, there must be meaningful human responsibility. AI can advise. Humans must remain accountable.
04
Children should not become an experiment for rapidly evolving artificial intelligence.
AI products used by minors deserve particularly strong safeguards. Technological progress should make childhood safer—not more dangerous.
05
AI can strengthen cybersecurity. It can also dramatically increase the capabilities of attackers.
America needs a national strategy to protect businesses, government networks, hospitals, utilities, financial infrastructure and citizens from increasingly automated attacks. We support stronger AI-security testing, threat sharing and defenses for critical infrastructure.
06
AI will make many workers dramatically more productive. It may also transform or eliminate entire categories of work. Pretending disruption will not happen is not a plan.
The AI economy should create widespread prosperity—not prosperity for only a few.
Rigorous evaluation of exceptionally capable systems before deployment into high-risk environments.
Strong safeguards against using advanced AI to meaningfully facilitate biological or chemical attacks.
Defenses against AI-assisted attacks on America's digital and physical infrastructure.
Meaningful human responsibility for appropriate high-consequence AI decisions.
Stronger protections for minors interacting with increasingly persuasive and autonomous AI systems.
Meaningful disclosure and accountability when powerful AI systems create substantial public risks.
Preparing Americans for rapid economic changes produced by automation.
Policies that allow American companies, entrepreneurs and researchers to continue leading global AI innovation.