Jai Ramaswamy is the chief legal and policy officer at Andreessen Horowitz, a venture capital firm that backs entrepreneurs in technology.
Last week, President Donald Trump issued an executive order calling for a national policy framework on artificial intelligence to bolster America’s leadership in technology. The executive order signals that the administration recognizes the need to clarify the confusing legal landscape that has emerged in AI, with an increasing number of complex state regulations and no federal guidance on how to govern this crucial emerging technology. Though the executive order is a welcome first step, it’s not enough. To solve this problem, Congress should take action.
While debate rages about the right way to govern AI, Americans should be able to agree on three truths. First, we are in an AI race against China — with China’s open-source AI models starting to exceed the capabilities of American-made tools — and the world is a safer, more prosperous place when the United States leads. Failure of U.S. leadership in AI will have enormous implications for our national security, economic competitiveness and global standing.
Second, technology start-ups have driven American economic productivity and technological superiority, from Ford Motor Company at the turn of the century to SpaceX more recently — and successful tech start-ups have an outsize impact on stock market returns. Start-ups drive innovation directly by innovating themselves and indirectly through competition with incumbents. We’re watching this pattern play out with AI: Between 2013 and 2023, more than 5,500 AI start-ups in the U.S. received funding, and numerous new cutting-edge AI products have come from start-ups. Recent developments by Google were spurred by competition in large language model development and advanced chip design. Start-ups continue to be a critical ingredient to ensure that the next era of technology is built in America.
Third, any effort to regulate AI must protect people while also safeguarding innovation. Policymakers can protect Americans from real harms — such as fraud or other crimes — and support a competitive, innovative AI market.
Consistent with these truths, the most important step that U.S. policymakers can take is passing a strong federal AI standard — one that positions America to lead, empowers start-ups to compete and creates a safe, trusted AI market for all Americans. Today, my firm, Andreessen Horowitz, a venture capital company that invests in and advocates for Little Tech, offered a road map for federal legislation to advance these principles.
Our proposed policy approach asks Congress to take meaningful action to ensure that AI cannot be used as a “get out of jail free” card for those who break the law. For example, if someone uses AI to commit fraud by impersonating a person or creating fraudulent materials, they have still committed fraud and should be held accountable. We also call on Congress to set the standard on protecting minors from AI-related harm, ensuring providers offer parents meaningful controls such as privacy and content settings, use limits, as well as requiring AI providers to develop and disclose protocols for detecting issues such as suicidal ideation or self-harm.
Congress should also invest in AI research and implement education programs that teach the next generation how to succeed in an AI-driven economy. The government should launch workforce development initiatives that provide AI training along with private sector partnerships that can lead to job opportunities. Congress should also modernize the successful but 80-year-old National Apprenticeship Act to be ready for the AI era by creating structured and streamlined standards that lead to AI careers.
Lawmakers should level the playing field for new entrants into AI. For instance, creating a new public entity — the National AI Competitiveness Institute — could help lower barriers to entry for entrepreneurs, small businesses, researchers and government agencies through public access to computational power, datasets and other tools.
America’s lack of uniform federal regulation for AI is damaging to start-ups and insufficient for consumers. Requiring two-to-three person teams to navigate a 50-state patchwork of AI laws — and have the same audit requirements and paperwork burdens as tech giants — does not make consumers safer. It creates a regulatory moat that will further advantage large, legacy companies that already possess the enormous resources required to train AI models, limiting the innovation that competition brings. Without a national standard, AI development will be concentrated in a few large companies that can afford to navigate inconsistent state laws and won’t face competitive pressure to be nimble, forward-thinking and consumer-friendly.
Moreover, the biggest beneficiary of the status quo is China. Models developed in China are already as effective and cheaper than American rivals and therefore provide a practical alternative for the millions of Americans who use AI. If we want to maintain American leadership in AI and all that it brings, it is imperative that Congress seize the moment and act now.
The post Trump seeks to lead on AI. Three truths are self-evident. appeared first on Washington Post.
