Charting America’s Course in the AI Age
When Chinese AI startup DeepSeek unveiled its groundbreaking new model earlier this year, it sent shockwaves through Silicon Valley and Washington alike. Deepseek’s breakthrough drew comparisons to Sputnik – serving as a wake-up call that America’s AI innovative dominance was no longer assured.
As this technology rapidly develops, governments across the world are facing mounting pressure to establish equitable frameworks that balance economic goals with privacy and national security interests. The President’s answer to this global race is his recently announced AI Action Plan, in accordance with the January executive order on Removing Barriers to American Leadership in AI. This new plan, which was implemented through three new executive orders, sets the stage for America’s competitive response to an era of digital infrastructure and algorithmic influence.
Global competition
After OpenAI’s release of ChatGPT in late 2022, China ramped up its investment in chatbot research and development. Its ability to respond so quickly with DeepSeek shocked U.S. companies and policymakers, forcing the realization that we are in a global race to lead AI innovation. In the time since DeepSeek’s announcement, many AI companies have emerged across the globe. Each is likely to have strengths and weaknesses, and robust competition will lead to a stronger overall AI ecosystem.
Global AI competition is intense. Without proper leadership, AI models could embed authoritarian values rather than democratic ones. With the White House’s new AI Action Plan, implementation should focus on bolstering America’s leadership and competitiveness in AI systems. It should also incentivize enhanced development of speciality AI tools to ensure diversity in the marketplace and continued success of U.S. leadership in the global AI race.
The AI Action Plan
The White House AI Action Plan directs the federal government to limit overly restrictive AI regulations at the state and federal level, avoiding compliance burdens associated with excessively burdensome state and regulatory laws. This is important partly for smaller companies and startups, who would rather spend their limited budgets on engineers instead of lawyers to address duplicative and conflicting legal compliance requirements. This plan also includes measures to boost AI infrastructure by fast tracking permit approvals to build data center infrastructure.
Considering government procurement has historically boosted technological innovation and development, another executive order directs the federal government to promote different AI models. This tech-neutral language helps different models compete, which encourages rapid AI development.
The AI Action Plan also encourages the development of open-weight/open-source AI models alongside the more typical closed-source models. These models have their own distinct shortcomings from closed source models. However, they also present significant benefits by allowing users to deploy the model as needed for their own equally innovative uses—a significant benefit, as open source software has illustrated for decades.
Addressing lingering division
While much of this new AI Action Plan is positive, one more controversial provision cautions against AI models promoting “wokeism.” Any executive order that instructs American AI developers to include or exclude particular viewpoints will inevitably be noted by authoritarian regimes to limit otherwise lawful conduct. Furthermore, knowing that models have been manipulated to promote or detract from specific information will reduce user trust in the neutrality and accuracy of AI results. This mistrust would apply not just to one application or system, but to every use of AI. In sum, unless users trust AI, it won’t be used, regardless of how many barriers to development are removed.
Conclusion
Policymakers looking to support U.S. AI must focus on securing both the technology’s material and policy needs. First, the U.S. must ensure that necessary infrastructure is in place, from electricity, energy, semiconductors, and data centers, as well as quality spectrum access and high speed computing. And second, policymakers must foster a policy environment that allows for continued innovation while balancing the novel privacy, security, and ethical concerns raised by this technology. This will not be easy, but the White House’s new AI Action Plan is a notable start.
