Artificial Intelligence,Rising Global Division,and Co-Bot Trolls: Group Researchers – We Got Some Work to Do!

Abstract

In my mid-career reflection, I want to highlight three research areas that require urgent attention. I themed these research opportunities as the ‘good’ (i.e., new AI methods changing how we can study groups), the ‘bad’ (the rise of hate, global division, and conflict), and the ‘ugly-ficial’ (working with vs against artificial agents). Looking into the next decade, my hope is that the ‘multi-disciplinary’ family that contributes to Small Group Research will work together to resolve some of our society’s most urgent problems.

Introduction

As I write this reflection in early 2026, I look at a world with increasing global international tensions (Henley et al., 2026; World Economic Forum, 2026). At the same time, Artificial Intelligence (AI) is fundamentally impacting the world of work (Bankins et al., 2024; Grimes et al., 2023; Islam & Clun, 2025). Many of us wonder if we woke up in a dystopian fever dream. These socio-political, economic and technological factors are ultimately impacting my stance on what group researchers should focus on over the next decade. I want to highlight three research areas that require urgent attention. I themed these research opportunities as the ‘good’ (i.e., new AI methods changing how we can study groups), the ‘bad’ (the rise of hate, global division, and conflict), and the ‘ugly-ficial’ (working with vs against artificial agents). Looking into the next decade, my hope is that the ‘multi-disciplinary’ family that contributes to Small Group Research (Emich et al., 2020) will work together to resolve some of our society’s most urgent problems (Kane, 2026).

The Good: Will AI Methods and Access to High-Resolution Data Fundamentally Change How We Study Dynamics?

Let’s start this essay with a positive outlook (before we get to the grim part). In our field, there has been a long debate that much of our body of knowledge around team dynamics is derived from static relationships (e.g., Cronin et al., 2011; Kozlowski, 2025; Lacerenza et al., 2025). I echo these concerns: If we truly want to move from a variable-centric to process-oriented theories of team dynamics (Kozlowski, 2025), we need methods that can track ‘what is going on’ in teams using a ‘high resolution’ lens (Klonek et al., 2019; Mathieu et al., 2022). Ideally, we want to use these methods ‘in the wild’, that is, by looking at teams operating in their fully situated context (Rosen et al., 2012). In these contexts, researchers can work with stakeholders and organisations who are truly hungry for scholarly advice and support (Klonek et al., 2020). Teams ‘in the wild’ can be emergency response teams (e.g., Schmutz et al., 2018), firefighters (e.g., Rico et al., 2025), or (to make things even more complicated) teams working in isolation and confinement (Gedik et al., 2023; Hagemann et al., 2025; Kozlowski, 2025). However, a key problem is that research in this context involves significant challenges when it comes to ‘measuring’ what ‘is going on’. First, research ‘in the wild’ should ideally be unobtrusive as we do not want to distract teams while they are trying to save lives or while dealing with an imminent crisis. Second, researchers want to measure multiple processes continuously as problems may occur due to communication breakdowns, lack of cohesion, bad leadership, or a zillion other things. Third, when going ‘into the wild’, researchers want to swiftly adjust their measures to cater for unanticipated events as these teams are embedded in a volatile, uncertain, complex, and ambiguous (VUCA) environment. We know that standard measurement designs are quite ineffective in dealing with these challenges. Yet access to ‘big data’ in combination with novel AI-based methods have the potential to substantially move our field forward (Kozlowski et al., 2016; Lehmann-Willenbrock, 2026; Luciano et al., 2018; Park, 2026; Woo et al., 2024). We are already seeing some applications of innovative machine-based methods (e.g., using text data, Klonek et al., 2025; Mathieu et al., 2022) and novel types of sensory data (e.g., Gedik et al., 2023; Lehmann-Willenbrock & Hung, 2024). For example, natural language processing (NLP) can be used to measure team processes from team communication transcripts (Mathieu et al., 2022). NLP can also be combined with data obtained from online communication platforms to index constructs like team burstiness (i.e., how much teams concentrate their communication during relatively contained time periods vs spreading it out over time, Riedl & Wooley, 2017) or diversity-related constructs (i.e., how much meanings conveyed by group members diverge from one another, Lix et al., 2022). My hope is that journals, beyond SGR, will give more space to these methodological innovations. More importantly, novel methods could be used to provide teams with more immediate process feedback which opens the space for the evaluation of team interventions and better practical support (Buengeler et al., 2017; Farh & Chen, 2018; Handke et al., 2022).

The Bad: Can Our Research Contribute Towards Understanding and Resolving Tensions at Micro, Local, and Global Levels?

The world is in a state of increased polarization (Iandoli et al., 2021; Van Bavel et al., 2021), with widespread online hate (Mak et al., 2024), an increase in armed conflict (International Committee of the Red Cross [ICRC], 2025), and rising geopolitical tensions (World Economic Forum, 2026). In many countries, we see efforts to dismantle and threaten democracies (International Institute for Democracy and Electoral Assistance, 2023). This begs the question how the family of disciplines that contribute to SGR (Emich et al., 2020) can provide insights and a meaningful impact to these severe problems that fracture our society (Kiesler, 2025). These questions and problems can be addressed from multiple disciplinary angles, including sociology, political sciences, communication science, and psychology, but at their heart, they involve conflict dynamics within and between groups which is the bread-and-butter of group researchers (e.g., O’Neill et al., 2018). To address issues like hate and aggression, our field must remain open towards novel methods, novel types of data, and study novel emergent phenomena. This type of research is not always easy (e.g., Klonek et al., 2023; Paletz et al., 2011), but we need to focus on emerging phenomena that are fundamentally impacting the world that we are living in (Lumineau et al., 2025).

The Ugly-Ficial: How Are Artificial Agents and (Co)bots Changing Team Dynamics? Is It Going to Get Ugly?

Only a couple of years ago, human-autonomy teams (HAT) research was anything but mainstream. Much was done within the human factors discipline using lab-based simulations (Glikson & Woolley, 2020; O’Neill et al., 2022). However, since the global release of generative AI chatbots (early 2023), these synthetic agents are increasingly shaping and transforming our ‘world of work’. An implicit assumption in the definition of teams is that they are comprised of ‘two or more [human] individuals . . .’ (e.g., Kozlowski & Ilgen, 2006, p. 79). Yet the inclusion of artificial agents is fundamentally changing team dynamics. So, can we just apply classic theories of team performance to human-AI/robot teams? Or should we rather develop fundamentally new theories to explain the specific changes that we see in these novel hybrid human-AI teams? In doing this, we must ask how attributes of artificial agents may impact individual, team, organisational and even societal outcomes. Again, I see the use of novel methods and new forms of collecting data as critical in answering these questions (Klonek & Parker, 2025; Klonek et al., forthcoming; Woo et al., 2024). Researchers also need to think how they can study AI-based phenomena in novel spaces (online, virtual, metaverse) and how these new conditions are shaping new phenomena (e.g., online shit storms, hate speech, manosphere etc.). Unfortunately, these phenomena have been severely neglected in SGR, which is why we need much more attention here (Kiesler, 2025). I have called this theme ugly-ficial (not only for the sake of the pop-cultural movie reference) but to call for a stronger focus on the ugly side of the ‘AI hype’. Synthetic actors who are contributing to group-based interactions might significantly accelerate the speed at which dynamics develop (e.g., viral spreading of misinformation, rumours, and online hate).

In sum, the next ten years (and beyond) should keep us all busy. Technological advances and AI allow for more high-resolution measurement of team process dynamics (the ‘good’). Yet we must spend more effort towards understanding and resolving tensions and conflicts in and between groups (the ‘bad’). Finally, we must understand how artificial agents might fundamentally shape team dynamics and how they contribute towards new phenomena that develop in groups (the ‘ugly-ficial’). Group researchers – we got some work to do!