U.S. DoD AI ambitions are broad and growing, but sourcing the necessary leading-edge chips to expand technology will bring challenges.
Growth in artificial intelligence (AI) technology represents the leading edge for industries from agriculture to social media. Additionally, the U.S. Department of Defense has ambitious plans to implement AI across all services. But, in addition to traditional concerns about security and ethics, chip supply chain issues add another layer of complications to maintaining pace with advancements in AI. Effective implementation of these leading-edge technologies will require conquering both administrative and acquisition hurdles to realize U.S. DoD AI ambitions.
Let’s look at some of the complications in successfully integrating AI up and down the Defense Department.
3 challenges facing ambitious U.S. DOD AI efforts
1. Bureaucracy and lack of centralized governance
A massive undertaking comprising more than 600 AI efforts is underway across the U.S. military, touching a wide variety of functions and services. But, advanced technology prompts the basic logistical issue of rolling it out across a large organization.
Air Force Magazine notes that in this enormous range of efforts, “some commands are just trying to get a handle on their data — what they have and how to format it for future use, for things as simple as optimizing a schedule.” Strong, clear leadership is necessary for driving adoption when such is the case.
The Pentagon acknowledges that streamlining and reducing bureaucracy are key to successful implementation. As of June 2022, it is standing up the Chief Data and Artificial Intelligence Office (CDAO) with the goal of linking multiple AI offices across the U.S. military forces. Such centralization of governance is one administrative strategy the DoD will need to prioritize to fully integrate AI.
2. Acquisition of talent and tech
Security is one of the Pentagon’s primary motivators for AI adoption, fueled by findings from The National Security Commission on Artificial Intelligence (NSCAI), which convened in 2019. The resulting report, published in March 2021, recommends advancing “the development of artificial intelligence, machine learning, and associated technologies to comprehensively address the national security and defense needs of the United States.” Therefore, not only are there hundreds of efforts in the pipeline, but also an adoption of artificial intelligence needs to be immediate and swift to protect our country from foreign threats.
The NSCAI acknowledges that AI depends on a stack of “interrelated elements” – two of which include the advanced technology and actual humans capable of running it. The acquisition of both admittedly have been a challenge, particularly with an accelerated timeline. The Secretary of Defense noted that to reach U.S. DoD AI goals, both acquisition of tech and talent needed to improve, with a need to “smarten up our sluggish pace of acquisition.”
When it comes to the technology, working to streamline acquisitions processes is only one solution. But identifying secure sources for electronic components will also be critical for moving forward with building leading-edge technology.
3. Leading-edge chips
In particular, one of the major issues driving the military’s slower pace of acquisition is the rapid development of the leading-edge chips that fuel AI. Without the most advanced chips, AI won’t function at the leading edge. These are currently not manufactured in the U.S., but, as we’ve written before, investment is building capacity.
Right now the Center for Strategic and International Studies reports that “Intel makes field-programmable gate arrays (FPGA), which are incorporated in AI systems, based on 10 nm design rules — a generation behind 7 nm.” But Intel won’t begin to produce 7nm chips in the U.S. until later this year. Meanwhile, the TSMC fab in Arizona will begin a 5nm node operation in 2024, but CSIS predicts that by that time, the leading edge chips will likely be 3nm.
Is this enough to stay on top of AI developments in coming years? Keeping pace with the leading edge of chips will create an ongoing challenge on top of administrative and logistical issues.
Chips will continue to fuel U.S. DoD AI ambitions, raising a host of security, ethics, and supply chain challenges for the Pentagon to tackle. The current strong administrative effort to organize the DoD around these ambitions will require smart sourcing to gain (and keep) the leading edge of AI technology.
Read more: