US May Be Losing the Race for Global AI Leadership

COMMENTARY

As a global leader in artificial intelligence (AI) innovation, the United States currently enjoys a significant competitive advantage economically, militarily, and technologically. To maintain this leadership, Congress and federal regulatory agencies must avoid complacency and recognize that our foreign competitors are diligently working to surpass us.

According to a recent report by the United Nations World Intellectual Property Organization, China filed more than 38,000 generative AI patents between 2014 and 2023, demonstrating its aggressive push to dominate AI advancement. This surge in patent activity underscores the critical need for the United States to accelerate its ongoing research and development efforts to maintain our global AI leadership.

To meet the demands of this moment, the US federal government must focus on three strategic AI research and development priorities that enhance our cyber resilience and national defense:

1. Clarifying the Definitions of Key AI-Related Terms

Ambiguity and inconsistency in AI terminology continue to pose significant risks and challenges. Without precise and consistent definitions for key AI-related terms like "open source AI," "AI security," and "trustworthy AI," our country faces challenges in ensuring the responsible use of AI. Focused research and development efforts toward closing this gap can help create a more comprehensive and nuanced taxonomy, mapping out potential varying definitions and detailing parameters. For example, standardizing the term "AI security" could not only help promote secure development practices but also address emerging challenges like prompt injections and ensuring visibility into AI usage across organizations to prevent data leaks.

By systematically analyzing and standardizing terminology through collaborative efforts across academia, industry, and government stakeholders, the United States can lay a foundation for developing robust strategies that protect our national security and promote cyber resilience.

2. Developing a Scientific Understanding of AI

A comprehensive scientific understanding of AI's limitations and capabilities is vital for assessing its impact and countering adversarial threats. The recent Private and Civil Liberties Oversight Board (PCLOB) AI Forum emphasized the importance of continuous research to measure AI's safety and reliability throughout its life cycle. To address this need, focused and strategically aligned research and development efforts are essential.

By developing ways to measure, test, audit, and evaluate AI capabilities, we can establish reliable methods to quantify AI's performance. These methods are critical for conducting accurate risk assessments and for our ability to continually improve AI systems and uses. Building a scientific understanding of AI, led by agencies like the National Institute of Standards and Technology (NIST) in collaboration with industry, academia, and other federal partners, will bolster our national and cybersecurity defenses by ensuring that AI systems are safe, reliable, and effective under various conditions and adversarial scenarios.

3. Developing AI Standards Adhering to US Mission and Values

Establishing AI standards that reflect our mission and values, such as transparency and accountability, is crucial for maintaining national security. The PCLOB AI Forum also highlighted the importance of developing clear standards that guide acceptable and non-acceptable uses of AI, particularly in the intelligence community.

Research and development can support this effort by identifying best practices and creating robust frameworks that ensure AI technologies are deployed ethically and securely. By focusing on these priorities, we can ensure that AI advancements align with American values and strengthen our national security and cybersecurity. This makes the development of AI standards an urgent and essential aspect of AI research and development.

Recognizing AI's pivotal role in national security, policymakers have already made significant investments in its advancement. To ensure these efforts remain focused and strategically aligned, Congress must consider targeted legislative measures while empowering federal agencies to facilitate coordination, promote public-private partnerships, and set actionable research and development priorities that encourage innovation within these established guidelines.

By prioritizing the clarification of key AI-related terms, building a scientific understanding of AI, and establishing AI standards that adhere to our mission and values, the United States can maintain the momentum of our advancements in AI and meet our cyber resilience and national defense imperatives.