Artificial intelligence (AI) is approaching a technological inflection point that irrevocably will alter the global perspective. In the coming years, the battlefield will be rife with automated and autonomous systems, and U.S. warfighters will be forced to keep pace with exponential advances in technology. How can the Department of Defense (DOD) integrate machine intelligence while mitigating the risks lethal autonomous weapon systems pose to humanity? The opportunity to tame the technological beast (AI) and save taxpayer dollars may be found in brain-computer interfaces (BCIs).
The next phase in the course of human evolution is on the horizon. Mankind will transition from being emotionally attached to our smartphones to forming an intellectual and physical symbiosis with computer systems. Instead of having touch screens as the choke point between the cerebral cortex and AI, human beings will mainline their dataflow using BCIs. These interfaces may be invasive devices—like cochlear ear implants for deaf patients and the neural arrays inserted into the brains of Parkinson’s disease patients—or non-invasive devices to be worn like helmets or skullcaps. Regardless of what forms they take, it is imperative that the U.S. military funnel resources toward the development and implementation of these symbiotic technologies.
The immediate warfighting benefits of BCI are most easily envisioned at the tactical level. The observe-orient-decide-act (OODA) loop, theorized by the late-Air Force Colonel John Boyd, is a widely accepted model of tactical decision-making. Traditionally applied to a pilot’s environmental acuity and consequential response in air warfare, the OODA loop is defined as:
Observation – the pilot observes, collecting data from all of his/her available senses.
Orientation – the pilot analyzes his/her situation and synthesizes the data that is available.
Decision – the pilot determines a course of action.
Action – the pilot realizes his/her decision through physical action.
Each phase of the OODA loop has an opportunity for reassessment as feedback is observed. Time is a parameter that often decides the outcome of combat, which is why speed of action is so important. In a tactical fight, combatants who make the best use of time progress through the OODA loop faster than their opponents and are likely to prevail.1 Though their enemy may have more powerful weapons and greater forces, savvy guerrillas often can outthink, outmaneuver, and defeat a superior opponent by turning inside their opponent’s OODA loop. This is why, in asymmetric conflicts, a technological overmatch does not always result in victory. Instead, it can be the cognitive overmatch that determines the outcome.
How can the U.S. military leverage BCIs to achieve tactical supremacy? The technological augmentation of the warfighter’s cognitive process—speeding up the OODA loop—will yield decisive results on the battlefield. An analysis of the hardware specifications and limitations of human cognitive systems (brains) shows that there is room for improvement. Neurons, the basic cellular unit of the brain, are loosely analogous to transistors in a computer. There is, however, one major difference. There are more than 100 billion neurons in the human brain, far exceeding the number of transistors in current computers, which equates to a great deal of processing power. For this reason, the speed and capacity of the human brain remain superior to present-day computer systems. Although highly specialized AI can beat human beings at board games (checkers, chess, go), AI-enhanced computers still are incapable of reading and interpreting a children’s storybook.
One of the major limitations of the brain, however, is that humans are output constrained. After integrating available senses to observe, humans orient and decide well, but they struggle to act quickly. This is in part because of the latency inherent in human gross motor responses. At best, the time between the reactive decision to respond to an observed stimulus and the voluntary movement instantiating that decision is no less than one full second.2 A one-second motor function delay may not sound long, but in modern combat the difference between life and death might be measured in fractions of a second. In addition to the latency issues associated with gross human movement, warfighters must have fine motor skills to manipulate current machine interfaces. Once a pilot, for example, has observed the environment and oriented to the threat, he or she may decide on a maneuver rather quickly, but it will take a complicated series of physical commands before the aircraft begins to respond to human input. Brain-computer interfaces will streamline this cognitive process and, essentially, press fast forward on the OODA loop.
Neuroergonomic approaches to interface design will allow warfighters to refine each phase of their OODA loops and leverage AI. When our brains speak directly to computers, the data collected by any netted sensors will be integrated into our sensory experience with an ease like that of vision from the human eye. A pilot would be able to observe through an electro-optical, infrared sensor mounted on the airframe as if it was a third eyeball. The pilot’s ability to orient would be augmented by AI—synthesizing large amounts of data to understand the environment. The symbiotic mind immediately will determine likely outcomes of potential actions and enable the human being to choose the best option. The most important component of the OODA loop—decision—should continue to be a human responsibility. This is a key ethical point in the development of cognitive technologies: humans should not give responsibility for decisions to AI.
Once oriented to their situations, humans will be rapidly and accurately informed to make the best possible decisions. Then, in lieu of the time-intensive process currently required to translate decisions into physical movement, the BCI will recognize the cortical potential of a decision and translate human intent into action. Instead of waiting for a pilot to smash buttons and yank a joystick, AI symbiotes will be able to translate human decisions and begin to maneuver the war machine before human fingers would have received the signal to move.
BCIs already can translate the readiness potential of decisional intent. In 2015, a team of engineers at the California Institute of Technology built prosthetic arms controlled by a BCI. Iterative improvements to the original design have been achieved across academia and at the Defense Advanced Research Projects Agency (DARPA).3 The prosthetic limb is equipped with a bundle of tactile sensors that are netted to an electronic array placed onto a paralyzed patient’s brain sensory and motor cortexes. The patient can detect and identify the slightest pressures applied to surfaces of the prosthetic arm and fingers. The patient not only moves the prosthesis as if it was a natural appendage but now has the tactile ability to feel a partner’s grasp. Neuroplasticity is the human brain’s amazing ability to adapt to varying modalities and inputs. Imagine if, instead of tactile sensors placed along a prosthesis, a pilot’s brain was connected directly to the advanced measurement devices and avionics of an aircraft. A connected pilot would be able to feel and manipulate the full capabilities of the airframe as if it were an extension of his or her own body. Employing weapon systems would be as natural as throwing a football, but faster and more coordinated. Our human experience is currently bounded by the spectrums of our sensory organs, but imagine if every component of an aircraft felt as organic as our eyes and feet. Simultaneously integrating sensors, manipulating weapon systems, and piloting high-performance vehicles would become extensions of human thought. This may be an abstract concept, but instead of just piloting an aircraft, a BCI-enabled aviator could feel what it is like to “be” an aircraft.4
As mind-blowing as such enhanced capabilities may seem, it will be the augmentation of human learning and memory that will have an even greater positive disruption. Although the formation of conscious memory within neurons and the symbolic representation of the human experience are unique to each individual, the merger of neuromorphic AI and high-fidelity BCIs could one day provide the Rosetta Stone to human thought.5 Creating modems that connect and translate human minds could open the door to unprecedented, collective capabilities, such as human memory storage, experiential knowledge transfer, and telepathic communication. This is why Elon Musk, CEO of Tesla and SpaceX, has created a new start-up called Neuralink, with the intent to develop and produce BCIs. Musk’s long-term goals include telepathic communication and combatting dangerous AI. Musk sees BCIs as a potential solution to the frightening Terminator or Matrix scenarios from Hollywood movies.6 If humans merge with the machines, then they can mitigate the risk of becoming slaves to them. Musk is not alone in the pursuit of BCI technology. Venture capitalist Bryan Johnson and Facebook CEO Mark Zuckerberg also have invested hundreds of millions of dollars into neurotechnology research.7
While tech titans like Musk, Johnson, and Zuckerberg recognize BCI as a potentially lucrative industry, how could a DOD investment in BCI save money? The answer to that question lies in reduced overall training and operating costs and improved scalability. Future military personnel, with cognitive augmentation enabled by BCIs, would be able to back-up their memories and, eventually, aggregate their memories into a common repository. Service members who have BCI-enabled memory storage would neither forget any aspect of their training nor need a refresher on rusty skills.8 With the addition of memory storage hardware, warfighters might never lose their edge. They would have the mental bandwidth to master multiple weapon systems and specialty skillsets without any memory-based degradation. The force could be comprised of multi-discipline experts who learn skills faster and require little to no sustainment training.
The next intellectual leap is to imagine the capabilities of 24/7 net-ready BCIs. With further development of telecommunications and global internet access, the introduction of cloud technology to augmented cognitive systems will reshape the military and the world. The crowd-sourcing of experiential knowledge will mean that, once one connected service member learns how to pilot the F-35, every member with access to the joint cloud could download this skill. The capability and experience of our nation’s most elite warfighters could be stored in the cloud and potentially uploaded to the brains of every recruit. If a service member tackles a new problem or innovates a worthwhile solution, then those emergent tactics, techniques, and procedures (TTPs) could be available to every other service member. The life-saving advantage of instantaneous TTP development and proliferation cannot be overstated. Anyone who has served in an environment seeded with improvised explosive devices (IEDs) knows that TTP dissemination saves lives. Multi-brain computing also would allow the members of a tactical unit to pool their collective perspectives to formulate a holistic view of the battlefield. Their collective consciousness and the supporting AI would be able to use every member’s observations to inform each decision-maker’s situational awareness. This kind of synergy would inform accurate and timely decision-making at every echelon of command. BCI-enabled connectivity dramatically would improve operational efficacy, minimize loss, and save dollars.
As with all access technologies, connecting BCIs to a network will pose ethical, health, and operational risks. If a BCI becomes a two-way street, then anti-tamper, spoofing, intercept, and brain control are all considerations that must be incorporated in the performance parameters. As these technologies, and the consequential evolution of humankind hurdle forward, DOD must have an oversight authority to mitigate the risks.
To bring these ideas to fruition, substantial research-and-development investments are required in the near term. The financial offsets will not materialize until the mid-term when cognitively augmented service members will be fielded across the U.S. military. The implementation of cloud-based repositories and net-ready BCIs will change the way we train and resource our military. In the long term, cognitive augmentation will result in greater capability and capacity while also supporting a net decrease in manning requirements and military spending.
If the entire force could remember every training event without experiencing degradation in performance, the need to train would diminish. Reducing total training time to maintain optimal readiness will minimize manpower costs and yield proportional reductions in weapon, equipment, and sustainment costs. If all members of the team are always sharp, they can train less often, operate their equipment less often, and spend less money on readiness.
If DOD achieved the ability to upload a compendium of experiential knowledge into a recruit’s neocortex through a globally accessible cloud, that would lower the cost of training for the entire military force and could make it more scalable. Much of DOD could resemble an inactive reserve, standing by to be called upon when needed. If fighter pilots and submariners could be trained by wirelessly downloading an app, then our inactive reserve service members could be the bulk of the force and would not be required to do much more than stay in good physical shape. They could live as healthy civilians until the nation needed their service. If large forces were needed to deploy for a purpose, the military could activate members of the reserve component, have them download specialized skills, and send them forward to complete their tasks with unprecedented speed and efficacy.
What kind of savings are possible with wide-scale application of BCIs? The DOD base budget submitted for fiscal year 2018 is roughly $521 billion.9 Personnel costs account for $135 billion. Operations and maintenance account for $197 billion.10 If the defense-wide application of BCI technologies could yield a 10 percent reduction in personnel, operations, and maintenance costs, then roughly $30 billion in annual savings is possible.
Since 11 September 2001, combatting and containing transnational terrorism has required massive expenditures. The malign influence of competitor nations on the international order threatens global commerce and stability. The U.S. national debt currently is nearly $20 trillion. The United States faces a resource dilemma that necessitates a reprioritization of DOD’s investment portfolio. Brain-computer interfaces and artificial intelligence offer potential paths toward tactical supremacy, defense-wide scalability, and significant cost savings for the U.S. military.
1. William S. Lind, Maneuver Warfare Handbook (Boulder, CO: Westview Press, 1985), 4-5.
2. Benjamin Blankertz, Laura Acqualagna, Sven Dähne, Stefan Haufe, Matthias Schultze-Kraft, Irene Sturm, Marija Uš?umlic, Markus A. Wenzel, Gabriel Curio, and Klaus-Robert Müller, “The Berlin Brain-Computer Interface: Progress Beyond Communication and Control,” Frontiers in Neuroscience, vol. 10, 2016, www.ncbi.nlm.nih.gov/pmc/articles/PMC5116473/.
3. “Neurotechnology Provides Near-Natural Sense of Touch,” DARPA, 11 September 2015, www.darpa.mil/news-events/2015-09-11.
4. Yuval Noah Harari, Homo Deus (New York, NY: Harper Collins), 2017, 356-364.
5. Nick Bostrom, Superintelligence (Oxford, UK: Oxford University Press, 2014), 57.
6. Liat Clark, “Elon Musk reveals more about his plan to merge man and machine with Neuralink,” Wired, 21 April 2017, www.wired.co.uk/article/elon-musk-neuralink.
7. Antonio Regalado, “The Entrepreneur with the $100 Million Plan to Link Brains to Computers” MIT Technology Review, 16 March 2017, www.technologyreview.com/s/603771/the-entrepreneur-with-the-100-million-plan-to-link-brains-to-computers/.
8. Robbin A. Miranda, William D. Casebeer, Amy M. Hein, Jack W. Judy, Eric P. Krotkov, Tracy L. Laabs, Justin E. Manzo, Kent G. Pankratz, Gill A. Pratt, Justin C. Sanchez, Douglas J. Weber, Tracey L. Wheeler, Geoffrey S.F. Ling, “DARPA-funded efforts in the development of novel brain–computer interface technologies,” Journal of Neuroscience Methods, Volume 244, 15 April 2015, www.sciencedirect.com/science/article/pii/S0165027014002702.
9. This does not account for Overseas Contingency Operation (OCO) funding which supports named operations like Operation Inherent Resolve.
10. Office of the Secretary of Defense Comptroller. “Defense Budget Overview: United States Department of Defense Fiscal Year 2018 Budget Request.” http://comptroller.defense.gov/Portals/45/Documents/defbudget/fy2018/fy2018_Budget_Request_Overview_Book.pdf.