The U.S. and its allies must do more to build an AI-capable workforce and transform the Defense Department’s culture to embrace artificial intelligence.
The future of warfare is changing at an accelerated rate. The battles of the future will have less to do with Cold War-era equipment (planes, tanks, ships) and everything to do with data, information processing and the effective deployment of machine intelligence. Next-generation warfare will be enabled by artificial intelligence, network centric and light-speed fast. The stakes are not just high -- they are staggering. The victor will hold the high ground with compounding first-mover advantage.
There’s no doubt about the ambitions of the Chinese Communist Party to lead the world in AI. Xi Jinping said China will be the world leader in AI by 2030, and he is directing national policies in support of this goal. Testifying before the Senate Committee on Armed Services recently, Eric Schmidt, chair of the National Security Commission on Artificial Intelligence, said he is “convinced that Chinese leadership in key technology areas is a national crisis that needs to be dealt with directly, now.” He told lawmakers that China could surpass the U.S. in AI within the next decade, and during questioning, Schmidt said he thought the U.S. is maybe one or two years ahead of China in certain applications of AI.
Given that thinnest of margins, the U.S. and its allies must do more to build an AI-capable workforce and transform the Defense Department’s culture to embrace AI. To accelerate the adoption of AI, I recommend a three-tiered approach.
First, the Pentagon should adopt an automation-first data science approach, which has been adopted across all industries, including cutting-edge finance. With this approach, data scientists can leverage automation across the entire AI lifecycle to test problem framing feasibility quickly, generate a robust first result and then layer in their expertise. Modern automated tools reduce the timeline of AI outcomes from months or years to days or weeks. Models leveraging open-source tools that provide the explainability and trustworthiness gives fledgling data scientists a safe path to automation. The alternative, and the status quo, is for DOD to hire significantly more PhD-level data scientists to create all AI solutions. However, with a shortage of U.S. STEM talent and the lure of high-tech AI startups, these federal data science jobs will remain unfilled, and the country will continue to fall behind.
Second, DOD should leverage enterprise AI platforms that enable a broader set of personnel (e.g. software developers, research analysts) with just a few days training to build AI solutions. Modern enterprise AI platforms leverage automation and provide a simple graphical user interface that allows people without deep experience in machine learning techniques and software programming to build a suite of predictive algorithms.
For example, a business analyst built a reliable AI solution to identify the factors that most closely predict success of candidates for special-forces training, aiding both the candidates and the military services. The beauty of automated machine learning is that it boosts productivity while freeing up under-staffed DOD data scientists to spend their time solving more complex problems, like developing next generation AI to combat cyber threats.
Third, the department must implement an AI capability building and culture-change program that is pervasive at all levels of the organization, from front-line troops to DOD leaders. Not everyone needs to be a data scientist, but everyone must know how and where AI can be deployed and how to leverage machine intelligence to meet mission requirements.
For example, warfighters must be trained to understand, integrate and implement the versatile power of AI -- the applications of which abound. In addition to more effectively putting the right personnel into the right training programs, AI can identify warfighters and veterans at risk for suicide, identify patterns to predict cybersecurity attacks and anticipate which tanks, trucks or helicopters need maintenance before they strand soldiers in the field. AI can also be used to more accurately forecast demand for fuel, ammunition or other supplies.
There is no doubt the United States must devote more resources to developing superior AI technology and to training more data scientists. Not doing so puts the country at great technological, economic, military and even personal risk. As we build these capabilities, it’s essential that personnel at all levels of the DOD develop a deeper understanding of how, when and where to use AI.