Artificial Intelligence (Formerly Emerging Technologies) –
“The global AI race will be won by whoever uses it most and best.” The 2023-24 Eisenhower School Artificial Intelligence (AI) Industry Study seminar – “Johnny Five” – internalized this message over the course of 45 engagements with government, industry, and academia in both the United States and Taiwan. Our key findings focus on innovation clusters, cyber-physical systems, data optimization, and keeping the U.S. edge in several key AI domains. Also, Johnny Five deliberately experimented with generative AI (GAI) tools to assist our work.
While AI hardly constitutes a traditional industry, the group managed to apply industry analysis tools to evaluate this fast-moving technology through the lens of competing ecosystems – one U.S.-led, and the other being developed by the People’s Republic of China (PRC) – and discovered intense but symbiotic competitive rivalry. At the heart lies the U.S.-PRC geopolitical contest, as the two superpowers pursue competition for AI supremacy, pouring massive investments into research and development, talent cultivation, and application.
A complex, global supply chain underpins the entire AI hardware-software stack, which runs on data, a specialized workforce, and computing power. The latter centers on the highly concentrated semiconductor industry, with U.S. firm NVIDIA leading AI chip design and Taiwan’s TSMC dominating the production side. U.S. investments resulting from the CHIPS and Science Act will take years to spur new output, leaving the United States and its allies vulnerable to any instability affecting the Indo-Pacific. These dynamics inform our four key findings.
1. Clusters can fast-track innovation but require concerted efforts by government, academia, and industry to create tech ecosystems, as seen in Silicon Valley, Pittsburgh, and Taiwan. Engaging with existing hubs can maximize the U.S. and DoD’s AI potential.
2. The world is transitioning from the information age to a cyber-physical age to be defined by the fusion of physical domains with data. This calls for intentional STEM workforce development and reintroducing a mentality of “tinkering” in the United States.
3. Data is paramount for AI. Models' utility reflects the quality, quantity, and security of data that feeds them. Techniques like multimodal sensing, digital twins, and synthetic data can optimize data, but ushing imperfect data or algorithms to market raises ethical concerns.
4. Washington and Beijing bring contrasting approaches and strengths to the AI race. The United States currently enjoys competitive advantages in four key AI areas – the AI stack, industry-academic partnerships, next-generation computing, and data optimization – but cannot take this lead for granted.
In addition to these findings, Johnny Five deliberately experimented with using GAI tools. Students found GAI was useful as an initial tool to accelerate learning of new content areas. But the technology appears to lack originality, generate repetitive content, and struggle with deep analytical tasks that demand human intuition and complex reasoning. For a resourcing exercise, however, GAI proved valuable for rapidly assessing service budgets against national priorities and identifying options for reallocating resources, creating more space for substantive dialogue and debate.
Our team recommends formally including AI in professional military education through hands-on practice and a culture shift. Johnny Five students now see themselves as informal AI ambassadors, ready to advocate for AI’s responsible use and diffusion throughout the national security establishment.
Read the report →