As military leaders, the true purpose behind a leadership discussion should be to criticize the process by which we make decisions and find ways to improve the process to increase efficiencies and innovation pertaining to warfighting capability.[1] For continued future military success, a discussion on military leadership needs to be applied to autonomous systems (AS), as these systems are the future of warfare execution in the terrestrial and space domain.
Leadership Assumptions and Terminology
A universal definition of leadership is impossible to establish, and there are various types of leaders: transformational, laissez-faire, charismatic, and transactional. Military officers should use a combination of these styles to effectively motivate and inspire those they are managing (the subordinate soldiers, sailors, Marines, artisans, etc.) General assumptions regarding leadership include that the leadership trait is driven by both a sense of morality and a sense of purpose independent of the ethical decision-making process; that military leadership decisions made to formulate and execute objectives are largely based on goal setting; and that individual internal and external influences drive the successful implementation of leadership.
Application-Specific and Scenario-Based Leadership
The application of sound leadership principles relies on the individual’s background and moral and ethical framework. In a dynamic military environment, leadership relies on the sense of judgement based on the operational setting.[2] Not every leadership style will work in every situation, therefore a developed framework tailoring leadership to the operational environment is critical.[3] For example, an officer assigned to a depot conducting maintenance may have one leadership style to create incentives and motivate the civilian aircraft artisans to strictly adhere to technical publication guidance on maintenance to produce high-quality work. Conversely, an officer attached to an expeditionary unit may have other methods to motivate or inspire Marines to produce consistently high-quality results. In either case, the leadership quality can be both independent of morality and driven by operational requirements, or strictly driven by a sense of morality and based on ethical principles, such as integrity. Additionally, the team dynamic can dictate or influence the leadership style that works best for a achieving a goal.[4] Leadership decisions made independently of an ethical process can still result in positive outcomes in professional and combat settings.
Non-Ethical Basis for Decision Making
A constant message to those who seek military leadership roles is to “do the right thing” and use a moral compass, as there is an assumed correlation between ethical behaviors and leadership as a trait. However, according to the Harvard Business Review, there is excessive pressure to reach unrealistic performance targets and that “unfettered goal setting can encourage people to make compromising choices in order to reach targets, especially if those targets seem unrealistic.” [5] Such pressure and goal setting can lead individuals in leadership positions to lie about progress or cut corners to achieve the end goal.
This pressure may negatively influence a leader’s ability to establish and manage effective risk decisions.[6] If the operational scenario or application-specific issue is too challenging, a leader may turn to a decision-making framework other than the ethic-based leadership decision process. “People motivated by specific, challenging goals adopt riskier strategies and choose riskier gambles than do those with less challenging or vague goals.”[7]
In establishing a framework where autonomous systems can execute sound leadership decisions, goal setting could lead to completely disregarding ethics as calculated risk increases.[8] For humans, “goal setting has been promoted as a powerful motivational tool, but substantial evidence demonstrates that in addition to motivating constructive effort, goal setting can induce unethical behavior.”[9] From the junior officer’s perspective, much of the work in operational, training, technical, and professional settings are driven by goals. As a staunch advocate for unmanned systems integration, a goal-based metric for artificial intelligence systems decision making should be met with apprehension when considering embedding leadership traits.
Autonomous Leadership and Unmanned Programmable Logic
Autonomous leadership refers to the latitude for making risk decisions outside of a specific chain of command or organization.[10] This leads to the roboethical question of “can leadership be programmed, and if so, should it?” If it can be programmed, it is a quantifiable element that can be mapped, coded, and subsequently embedded into an AS.[11] Additionally, from a military officer training perspective, further analysis on autonomous leadership principles for humans may apply to programmable leadership for AS. Ethical decision-making models for AI are heavily researched, and finding a model that accounts for both human- and AS-based leadership decisions will aid in future warfighting operations.[12] In addition to qualitative-based analytics, quantitative mathematical tools used to analyze risk decisions for AS could also evaluate human-based leadership principles and decisions.[13]
External Variables and Drivers Impacting Leadership Success
Rephrase the question of “what makes a good leader?” into “what is the motivation or what are the influences that make someone want to be a leader?” Are there inherent characteristics or motivations that could predict success and are thus desirable for a junior officer?
Quantifiable indicators or characteristics could be past successes with other squadrons, units, or companies; fitness reports; award citations; or other quantitative metrics used to evaluate performance. There are also critically important qualities that make a good leader—good moral character, integrity, and judgement— but that are not considered metrics for future success. Since the quality of leadership is continuously refined and relies solely on an external point of view to determine whether it is “good” or not, our success as junior officers is dependent on how others view our judgment and leadership abilities—often through trial and error situations.[14]
External variables impacting the success and motivation of junior officers to lead include perceived pressure for mission execution, a sense of competitiveness in the workplace, an individual’s own sense of morality and judgement, and the individual’s perception within or outside a command or environment. It is crucial to consider these external factors: an AS has no motivations to be a leader except for what the coding dictates.
It is important to carefully develop a logical flow for machines and systems to truly operate as autonomous entities.[15] The leadership trait is complex and needs to be carefully “mapped” and analyzed for areas of conflict to prevent an AS from improperly executing the mapped logic. This improper execution could lead to enhanced risk behaviors, and potentially, a complete disregard for human life as quantitative metrics alone are used to meet specific programmed tasks or goals. It is recommended that humans remain in the loop for AS, and the human-machine interface is preserved during all applications of AS in military functions.
Conclusions
Analyzing how military officers conduct a logical thought flow based on an ethics-driven leadership style can identify areas of improvement when applying these principles to a programmable logic for autonomous systems. An essential question to ask when considering leadership as an embeddable quality in machines is “does the external perception of leadership still carry?” If a human’s military leadership quality is autonomous and there is latitude to make decisions outside of perception, should the same quality ever be extended to truly autonomous systems? If the human in question is even remotely reliant on perceptions as an external factor that can influence their making a potentially poor decision, what realistic tasks can an autonomous system be required or expected to execute? By qualitatively and quantitatively reviewing leadership as a trait unique to humans, researchers may be presented with an opportunity to ultimately refine and/or improve a military autonomous system’s decision process.
Endnotes
[1] R. Dodge, J. Dwyer, S. Witzeman, S. Neylon, & S. Taylor, (2017) “The role of leadership in innovation,” Research-Technology Management, 60(3), (2017): 22–29, doi: 10.1080/08956308.2017.1301000.
[2] P. R. Bollinger, “Military leadership in combat,” doctoral dissertation, Indiana University of Pennsylvania, 2013, https://pdfs.semanticscholar.org/55e5/42f4b3972e12237bf399cfb1309bfaac985e.pdf
[3] S. Harper, “The leader coach.”
[4] D. Solow, S. Piderit, A. Burnetas, & C. Leenawong, Mathematical models for studying the value of motivational leadership in teams. Computational & Mathematical Organization Theory, 11(5), (2005), doi: 10.1007/s10588-005-1752-0. Retrieved from https://link.springer.com/article/10.1007/s10588-005-1752-0.
[5] R. H. Bishop, “Decision-making using mathematics,” Journal of the Scholarship of Teaching and Learning for Christians in Higher Education, 4(1), (2010): 13–22. doi: 10.31380/sotlched.4.1.13.
[6] L. D. Ordonez, M. E. Schweitzer, A. D. Galinsky, & M. H. Bazerman, “Goals gone wild: The systematic side effects of over-prescribing goal setting” (Working Paper No. 09-083), Harvard Business School, 2009, http://www.hbs.edu/faculty/Publication%20Files/09-083.pdf.
[7] Ordonez et al., “Goals gone wild.”
[8] F. Alaeiri, “Ethics in social autonomous robots: Decision-making, transparency, and trust,” doctoral dissertation, University of Ottawa, Ottawa, Canada, 2018, https://ruor.uottawa.ca/bitstream/10393/37941/1/Alaieri_Fahad_2018_thesis.pdf; F. Alalieri & A. Vellino, “A decision-making model for ethical ro(bots),” 2017, doi: 10.1109/IRIS.2017.8250122
https://www.researchgate.net/publication/320238179_A_Decision_Making_Model_for_E thical_Robots; and S. Fourtane, “Roboethics: The human ethics applied to robots,” Interesting Engineering, 22 September 2019, https://interestingengineering.com/roboethics-the-human-ethics-applied-to-robots.
[9] Ordonez et al., “Goals gone wild.”
[10] M. Stevenson, “Autonomous leadership and ways to encourage autonomy in the workplace,” H.R. Exchange Network, May 8, 2018, https://www.hrexchangenetwork.com/hr-talent- management/articles/autonomous-leadership-and-ways-to-encourage.
[11] J. Torreson, A review of future and ethical perspectives of robotics and AI, Frontiers in Robotics and AI, January 2018, doi: 10.3389/frobt.2017.00075.
[12] N. Moradinasab, R. Soofifard & G. R. Asili, “Developing an effective mathematical model for leadership styles selection by using fuzzy logic: a case on RIPI HR characteristics,” International Journal of Productivity and Quality Management, Inderscience Enterprises Ltd, 19(4), (2016): 466–484, https://ideas.repec.org/a/ids/ijpqma/v19y2016i4p466-484.html; Torreson, J. (2018). A review of future and ethical perspectives of robotics.
[13] Alalieri, “A decision-making model”; Bishop, “Decision-making using mathematics”; Moradinasab et al. “Developing an effective mathematical model.”
[14] Bollinger, “Military leadership in combat.”
[15] R. Carucci, “Why ethical people make unethical choices,” Harvard Business Review, 16 December 2016, https://hbr.org/2016/12/why-ethical-people-make-unethical-choices; D. M. Messick & M. H. Bazerman, “Ethical leadership and the psychology of decision making,” Sloan Management Review, 37, (1996): 8-9.
