Synthetic Authority: Speculating the Future of Leadership in the Age of Human-Autonomy Teams

Abstract

The dawn of automation, often termed the fourth industrial revolution, is ushering several new waves of societal change. As existing scholarship demonstrates, considerations of human-autonomy teams across varied contexts are becoming crucial and vital. A critical arena of investigation that is currently understudied is the development of leadership in human-autonomy teams and the associated factors that impact emergent leadership patterns. While existing scholarship points to key leadership traits on the human side that will be needed to manage human-autonomy teams, further research is needed to fully visualize the efect of diferent leadership paradigms on the collaborative spirit of human-autonomy teams. Motivated by conceptual precursors, this position paper proposes a speculative design-based investigation aimed at understanding the expectations and perceptions around leadership in the context of human-autonomy teams. The goal of this investigation is to gain insights into the adaptions in workflows, enhancements needed in competencies for both human and non-human actors, novel task management and delegation routines, and contextual considerations that characterize the way in which leadership patterns can be conceptualized for human-autonomy teams. Further, through a critical turn, investigating the ethical and larger societal ramifications of such models is also another essential goal this proposed work aims to achieve. The ultimate aim of this work is to pave the way toward design recommendations and policy considerations associated with envisioning the construct of leadership in the human-autonomy assemblage.

1. Introduction

The rising ubiquity of Artificial Intelligence (AI) is causing widespread change and reorganization of existing social practices and norms [ 1 ]. Such changes are implicitly and explicitly reshaping day-to-day social functions and lived experiences of people across many diferent contexts [ 2 ]. Autonomy can be found in varied scales and scopes, ranging from being a mere tool that in the form of chatbots can aid information and knowledge discovery on websites, to being a more sentient teammate that can act, reason, and collaborate nearly at par with nEvelop-O (N. J. McNeese) (N. J. McNeese) mate, it becomes important to understand the impact of such collaborative endeavors on the level of cohesion,

Several studies have indicated diferent antecedents of

trust and reliance on autonomous entities that can play a pivotal role in sustaining human-autonomy teams [6].

Thus, leadership and management of such teams have thus become vital considerations [7]. The question of leadership is particularly essential in order to understand the manifestation and concentration

seen widespread exploration [13], it remains to be understood how concepts from human-human teaming evolve are named as the most widely studied and employed with the infusion of autonomous agents in the teaming models [20]. Leadership as these models theorize, has dynamic [14]. While some conceptual and empirical in- been investigated in terms of behavioral attributes that vestigations have been proposed and conducted, these can be critical in encouraging growth and enrichment of remain limited in their conceptual and contextual scope the overall organizational climate [21]. Empirical investiand relevance [15, 16]. The key gaps that exist revolve gations that build upon the findings of these models have around the following. (1) A deeper understanding of how indicated how leadership attributes can impact leaderthe notion of leadership (particularly from a structural follower relations, collective and individual morals, team perspective) may evolve to encompass the impact and in- and individual performance and essentially demonstrate lfuence of autonomous entities on teaming dynamics (2) the essence of leadership as the binding thread that deperceptions and expectations around autonomous leaders ifnes and holds a team together [ 22, 23]. (3) the impact of emergent leadership models on teaming outcomes, particularly trust. To address some of these 2.2. Leadership and Technology crucial gaps, a speculative design initiative is proposed to explore thoughts and rationales to capture critical nu- Leadership is a highly fluid and dynamic notion, with a ances associated with the construct of leadership. The variety of factors impacting such fluctuations [ 7]. Furmotive of using a speculative design approach is driven by ther, in this regard, with the advent and prolific adoption prior research that highlights the power of this method- of Information and communication technology (ICTs), ological paradigm to capture desires, beliefs, and values the notion of leadership needs to change to accommoin a futuristic context, thereby inspiring perspectives and date the presence and influence of non-human actors. guidelines that can impact the conceptualization of fu- A seminal synthesis of scholarly visions in this regard ture technical artifacts [17]. The contextual stage for this was provided by Larson and DeChurch (2020) [14]. This study shall encompass several domains, as the goal will be endeavor posited four key models highlighting various to understand the phenomenon of leadership and its im- ways in which technology can mediate and impact leadpact on Human-autonomy teams, instead of binding it to ership. The key takeaway from this exploration is that a specific contextual premise. The aim of conducting this the dynamic between technology and leadership entails exploration is to provide insights into the way in which where and how the technology is positioned, and how it leadership models particularly in the human-autonomy mediates and impacts the team dynamic through various teaming frontier can be more rigorously conceptualized afective and cognitive states (such as trust, cohesion, and the inferences, associations unpacked can further and shared mental models) which ultimately impact the inform the way in which autonomous agents factor into way in which leadership is practiced. Building on adapthe teaming dynamic. The expected outcomes include a tive structuration theory, a novel theoretical proposition deeper understanding of how the inclusion of autonomy postulated the interplay between leadership and ICTs, in the teaming dynamic calls for the need to adapt lead- mutually shaping one another, that states while technical ership and management routines and the way in which mediums impact the way in which leadership is strucsuch adaptions are perceived to impact the competencies tured, the style of leadership also, in turn, impacts the needed to sustain such teams. nature of appropriation of such channels that mediate and impact group processes [24].

In this regard, a major line of inquiry has primarily 2. Related Work focused on the way in which the emergence of virtual communication formats (such as email and groupware 2.1. Leadership in Organizational Studies systems) impacts leadership processes and the adaptaLeadership is an essential facet of any group-linked en- tions that are needed to acclimatize with the change in deavor [10]. Leadership is critical for the success, sur- modality [25, 26]. Competencies in the form of motivavival, and sustenance of any team-driven activity [18]. tional traits, behavioral characteristics, and empathetic Leadership is primarily crucial from the point of view stances that are needed to succeed in such leadership of establishing structure with entails creating routines roles have also been explored [27]. Contextual mediation and processes needed to manage and maintain order in a and the impact of factors such as the efect of the COVIDgroup. From the point of organizational studies, leader- 19 pandemic on workplace regimes and the influence ship has been investigated through various theoretical of situational needs on leadership styles are also critical perspectives to better situate the demands and needs as- lines of scholarship that have been investigated in this sociated with the leadership function [19]. Especially in context [28, 29]. While this line of scholarship continthis regard, diferent models of leadership (transactional, ues to grow in momentum, another key arena wherein transformational, autocratic, democratic, charismatic) leadership and technology intertwine is the frontier of human-autonomy teams. 2.3. The Rise of Human-Autonomy the impact and needed skill adaptation at the human Teaming level in order to incorporate, manage and negotiate the efects, and implications of artificial intelligence Before detailing the emergent stream of work associ- usage while also balancing team efectiveness and ated with leadership in human-autonomy teams, a brief performance [16]. In this study, which is most closely overview of the scholarship in the human-autonomy linked with this proposed endeavor, the goal has been teaming arena is warranted to set the stage for this pro- to highlight the skills and required characteristics posed study. Expanding the notion of human-human needed by human leaders within the military context teams, human-autonomy teams are defined as interde- to lead human-autonomy teams. While the insights pendent entities, collectively engaged in a shared vision, are critical, it does not fully capture the way in which purpose, and goal [5]. Pertinent in such discussions is leadership models impact team dynamics. It does the consideration of expectations and characteristics that not elucidate on structural perceptions of leadership, would be desired in autonomous teammates [30]. Essen- capturing the way in which workflows, roles, and task tial in such cases are the considerations of shared mental management will undergo potential changes. It does models that develop and are critical for such teams to not factor in the perceptions of the impact of leadership operate and thrive [31]. Agent characteristics and role on team dynamics and it does not capture nuances in the team dynamic can be crucial factors mediating the about perceptions around the viability of envisioning way in which such teaming endeavors are sustained [32]. autonomous leaders. Further, it is limited in context Studies have shown how perceptions of AI teammates since it only focuses on the case of military leadership. can impact the overall task efectiveness and interdepen- All these aforementioned limitations are the gaps the dence between teammates [33]. In aggregate, all these proposed investigation envisions to fulfill. Below, we investigations highlight key strands of scholarship that present a set of wide-ranging research questions that are demonstrate the growing need to further develop this meant to provide the foundations for future research and domain of investigation. Adding to this eclectic collage also aim to elicit further conversations around leadership of ideas and concepts, considerations around leadership and human-autonomy teams. can be essential to gain a deeper sense of the way in which team management, orchestration, and distribution of tasks across human and non-human actors can 3.1. Research Questions add another layer of conceptual nuance that presently is under-explored.

3. Research Framing: Leadership and Human-Autonomy Teams

Leadership in the context of human-autonomy settings involves several conceptual considerations. Several questions surrounding the impact and intertwining efect of autonomy also become critical [ 34 ]. A key line of inquiry in this regard has explored how extant literature in human-robot collaboration, positions robots with respect to human actors when elaborating on collaboration, taking a more network view on leadership to understand the embeddedness of autonomous agents [ 35 ]. Further, [15] posited how leadership can be envisioned in relation to resource sharing and management, drawing on literature from leadership investigations under the umbrella of organizational studies. This article while visionary does not capture the relational dynamics between human and non-human actors, as it takes an information exchange perspective to leadership. This paper calls for a deeper need for empirical investigations, which motivates the research plan proposed in this narrative.

Related explorations have considered humanautonomy teaming in contextual settings, investigating 1. How will diferent leadership styles impact team dynamics in human-autonomy teams? a) How might they impact workflows? b) How might they impact human autonomy and agency? c) How might they be characterized by contextual attributes? 2. What autonomous agent characteristics might impact and be needed to sustain leadership in human-autonomy teams? a) What characteristics may be expected and desired for an autonomous leader? 3. What critical considerations and concerns may arise when construing leadership for humanagent teams?

a) How might it influence ethical visions?

4. Proposed Methodological

Approach – Speculative Design

Although human-autonomy teams are likely going to ifnd widespread use in the future [ 16], it is still a nascent ifeld today, thus the goal of the motivating research questions stated above is to tap into the way people envision future use cases around leadership in human-autonomy teams. This is similar in motive to speculative design- each for 1 hour. Each workshop will have approximately based approaches, which have found frequent use to con- five to six participants. The workshops are envisioned ceptualize everyday use and appropriation of automated to be conducted as semi-structured interviews with the artifacts such as robots [17]. However, such approaches participants and will also involve some co-design activiare yet to be applied in the context of human-autonomy ties. At this time, we expect to conduct these workshops teams. Given that such techniques provide an efective remotely (to aford geographical flexibility in terms of approach to elicit considerations to ideate and deliberate workshop attendance). Additional details such as comregarding a futuristic context, it serves as an eficient pensation will be finalized based on recruitment status methodological tool to be employed for achieving the and as the plan of the design activities (e.g., the time goals associated with this investigation. needed and type of activities) develops throughout the

Design fiction-based approaches serve as one of the course of this exploration. key methodological paradigms for materializing speculative design [ 36 ]. Design fiction revolves around generating insights and spaces for discussion about technology 5. Expected Contributions and rather than focusing on finished products [ 37 ]. The cen- Future Extensions tral definition of design fiction stands at – “ the deliberate use of diegetic prototypes to suspend disbelief about change”. The expected contributions are to provide more insights In essence design fiction involves marrying perspectives into the way in which leadership considerations will imabout the product and the context envisioned through pact the structure and workflows of human-autonomy the story world. Particularly, the aim will be to use the teams. Secondly, the goal will be to elicit considerations method of futuristic autobiographies (FAB) that provides and apprehensions around the use of automated leadera method of implementing the diegesis stated by design ship [ 41 ]. Further, from a methodological standpoint, the ifction [ 38 ]. use of speculative design serves as a novel exploratory framework to investigate leadership in human-autonomy 4.1. Futuristic Autobiographies teams. Finally, the goal will be to elucidate how existing models that have been stated in the context of leadership (without the inclusion of autonomous teammates) will need to be adapted to suit the change in the relational fabric induced by the involvement of autonomous agents.

Based on the insights collected, future explorations will embark on conducting experimental design studies to understand perceptions of leadership models and how these are impacted by agent features and roles in the teaming dynamic. Subsequently, survey analysis shall be conducted to better understand expectations and idealized perspectives around autonomous leaders, akin to the synergistic visions that have explored attributes of ideal artificial teammates [ 30]. The long-term contributory goal is to provide design recommendations for constructing and conceptualizing autonomous systems and to also robustly outline policy considerations to enhance the experience of humans operating with autonomous agents in teaming contexts [ 42 ].

This method allows for embedding the user or study participant as a central character in the story world presented. Materialized in the form of fictional autobiographic experiences that also help to elicit values, positions, and expectations of the stakeholders involved. The key features of this method include: (1) Expeditious – sessions are usually 30 to 60 minutes in total and typically involve discussion on 2-3 themes; (2) Open-ended – scenarios are developed with some scope of discussion, yet given the endpoint is carved out, there is a guiding trajectory for participants to follow as they engage as central characters (3) Captivating content – the content or themes explored are developed with diverse context and experiences in mind (4) Values and ethics – centering themes such that they trigger discussions around values and ethics. Further, this method can be seen as an extension of qualitative interviewing and thus can be analyzed through constructivist grounded theory [ 39 ]. 4.2. Plan for Data Collection and Analysis The goal of this exploratory study is to lay the foundations and to build a more robust understanding of the issues as outlined in the research motivation. The population pool will engage those with some familiarity with technology and AI. Participants will be recruited by word of the mouth approach, through personal contacts of researchers, and through social media groups [ 40 ].

Presently, we plan on conducting two workshop sessions, humans’ perception of machines as teammates with [19] R. Hogan, R. B. Kaiser, What we know about leada systematic literature review (2022). ership, Review of general psychology 9 (2005) [4] N. J. McNeese, M. Demir, N. J. Cooke, C. Myers, 169–180.

Teaming with a synthetic teammate: Insights into [20] T. S. Nanjundeswaraswamy, D. R. Swamy, Leadhuman-autonomy teaming, Human factors 60 ership styles, Advances in management 7 (2014) (2018) 262–273. 57. [5] T. O’Neill, N. McNeese, A. Barron, B. Schelble, [21] H. Tohidi, M. M. Jabbari, Organizational culture and Human–autonomy teaming: A review and anal- leadership, Procedia-Social and Behavioral Sciences ysis of the empirical literature, Human factors 64 31 (2012) 856–860.

(2022) 904–938. [22] B. Omolayo, Efect of leadership style on job-related [6] N. Ezer, S. Bruni, Y. Cai, S. J. Hepenstal, C. A. Miller, tension and psychological sense of community in D. D. Schmorrow, Trust engineering for human-ai work organizations: A case study of four organizateams, in: Proceedings of the Human Factors and tions in lagos state, nigeria, Bangladesh e-Journal Ergonomics Society Annual Meeting, volume 63, of Sociology 4 (2007) 30–37.

SAGE Publications Sage CA: Los Angeles, CA, 2019, [23] Z. R. Mulla, V. R. Krishnan, Transformational leadpp. 322–326. ership: Do the leader’s morals matter and do the [7] D. Greenfield, The enactment of dynamic leader- follower’s morals change?, Journal of Human Valship, Leadership in Health services (2007). ues 17 (2011) 129–143. [8] J.-W. Hong, With great power comes great respon- [24] B. J. Avolio, S. Kahai, G. E. Dodge, E-leadership: sibility: inquiry into the social roles and the power Implications for theory, research, and practice, The dynamics in human-ai interactions, Journal of Con- leadership quarterly 11 (2000) 615–668. trol and Decision 9 (2022) 347–354. [25] J. E. Hoch, S. W. Kozlowski, Leading virtual teams: [9] S. D. Ramchurn, S. Stein, N. R. Jennings, Trust- Hierarchical leadership, structural supports, and worthy human-ai partnerships, Iscience 24 (2021) shared team leadership., Journal of applied psychol102891. ogy 99 (2014) 390. [10] J. M. Burns, Leadership, Open Road Media, 2012. [26] K. Mehtab, A. ur Rehman, S. Ishfaq, R. A. Jamil, [11] C. Flathmann, B. G. Schelble, R. Zhang, N. J. Mc- Virtual leadership: A review paper, Social Sciences Neese, Modeling and guiding the creation of eth- 8 (2017) S1. ical human-ai teams, in: Proceedings of the 2021 [27] C. Wiradendi Wolor, S. Solikhah, N. F. FidhyalAAAI/ACM Conference on AI, Ethics, and Society, lah, D. P. Lestari, Efectiveness of e-training, e2021, pp. 469–479. leadership, and work life balance on employee per[12] S. Sankaran, C. Zhang, M. Gutierrez Lopez, formance during covid-19, Journal of Asian Finance, K. Väänänen, Respecting human autonomy Economics and Business 7 (2020). through human-centered ai, in: Proceedings of [28] A. V. Roman, M. Van Wart, X. Wang, C. Liu, S. Kim, the 11th Nordic Conference on Human-Computer A. McCarthy, Defining e-leadership as competence Interaction: Shaping Experiences, Shaping Society, in ict-mediated communications: an exploratory 2020, pp. 1–3. assessment, Public Administration Review 79 (2019) [13] R. J. Vender, Leadership: an overview, Oficial jour- 853–866.

nal of the American College of Gastroenterology| [29] F. Contreras, E. Baykal, G. Abid, E-leadership and ACG 110 (2015) 362–367. teleworking in times of covid-19 and beyond: What [14] L. Larson, L. A. DeChurch, Leading teams in the we know and where do we go, Frontiers in psycholdigital age: Four perspectives on technology and ogy 11 (2020) 590271. what they mean for leading teams, The leadership [30] R. Zhang, N. J. McNeese, G. Freeman, G. Musick, ” quarterly 31 (2020) 101377. an ideal human” expectations of ai teammates in [15] C. Flathmann, B. G. Schelble, N. J. McNeese, Fos- human-ai teaming, Proceedings of the ACM on tering human-agent team leadership by leveraging Human-Computer Interaction 4 (2021) 1–25. human teaming principles, in: 2021 IEEE 2nd Inter- [31] R. W. Andrews, J. M. Lilly, D. Srivastava, K. M. national Conference on Human-Machine Systems Feigh, The role of shared mental models in human(ICHMS), IEEE, 2021, pp. 1–6. ai teams: a theoretical review, Theoretical Issues [16] A. Pierre, Leadership of Human-Machine Teams in Ergonomics Science (2022) 1–47. in Military Environments: An Exploratory Frame- [32] D. Siemon, Elaborating team roles for artificial work, Ph.D. thesis, Regent University, 2022. intelligence-based teammates in human-ai collabo[17] J. Auger, Speculative design: crafting the specula- ration, Group Decision and Negotiation 31 (2022) tion, Digital Creativity 24 (2013) 11–35. 871–912. [18] J. Gardner, On leadership, Simon and Schuster, 1993. [33] B. G. Schelble, C. Flathmann, N. J. McNeese,

References

[1] E. A. Nabila , S. Santoso , Y. Muhtadi , B. Tjahjono , Artificial intelligence robots and revolutionizing society in terms of technology, innovation, work and power , IAIC Transactions on Sustainable Digital Innovation (ITSDI) 3 ( 2021 ) 46 - 52 .

[2] S. Pink , M. Ruckenstein , M. Berg , D. Lupton , Everyday automation: Setting a research agenda ( 2022 ).

[3] J. Rix , From tools to teammates: Conceptualizing T. O'Neill , R. Pak , M. Namara, Investigating the effects of perceived teammate artificiality on human performance and cognition , International Journal of Human-Computer Interaction ( 2022 ) 1 - 16 .

[34] A. M. Smith , M. Green , Artificial intelligence and the role of leadership , Journal of Leadership Studies 12 ( 2018 ) 85 - 87 .

[35] C.-Y. Tsai , J. D. Marshall , A. Choudhury , A. Serban , Y. T.-Y. Hou, M. F. Jung , S. D. Dionne , F. J. Yammarino , Human-robot collaboration: A multilevel and integrated leadership framework , The Leadership Quarterly 33 ( 2022 ) 101594 .

[36] T. J. Tanenbaum , K. Tanenbaum , R. Wakkary , Design fictions , in: Proceedings of the Sixth International Conference on Tangible, Embedded and Embodied Interaction , 2012 , pp. 347 - 350 .

[37] J. Lindley , P. Coulton , Back to the future: 10 years of design fiction , in: Proceedings of the 2015 British HCI conference , 2015 , pp. 210 - 211 .

[38] E. Cheon , N. M. Su , Futuristic autobiographies: Weaving participant narratives to elicit values around robots , in: Proceedings of the 2018 ACM/IEEE International Conference on HumanRobot Interaction , 2018 , pp. 388 - 397 .

[39] K. Charmaz , L. Belgrave , et al., Qualitative interviewing and grounded theory analysis, The SAGE handbook of interview research: The complexity of the craft 2 ( 2012 ) 347 - 365 .

[40] O. C. Robinson , Sampling in interview-based qualitative research: A theoretical and practical guide , Qualitative research in psychology 11 ( 2014 ) 25 - 41 .

[41] P. Feshchenko , Algorithmic leadership and algorithmic management: a systematic literature review ( 2021 ).

[42] J. Qadir , M. Q. Islam , A. Al-Fuqaha , Toward accountable human-centered ai: rationale and promising directions , Journal of Information , Communication and Ethics in Society ( 2022 ).