Are warmth and competence of Generative AI relevant as teachers determine their collaborative potential?

• In Event: The impact of generative artificial intelligence (AI) on students and teachers

Proposal

Introduction
The rise of educational technologies (EdTech) has transformed the pedagogical landscape, introducing digital tools aimed at supporting teachers and enhancing student outcomes (Nye, 1997; Winn, 2002; Januszewski & Molenda, 2013; Selwyn, 2022). One notable shift has been the expectation for teachers to collaborate with their 'virtual counterparts' rather than merely use them (Urlaub & Dessein, 2022; Kaplan-Rakowski et al., 2023; Kim, 2023). This transition from using to collaborating with technology is underpinned by the recognition that teacher collaboration is a key driver of educational success (Vaughn & Baker, 2008; Meirink et al., 2010; Vangrieken et al., 2015). While collaboration among humans hinges on expertise and emotiveness (Forte & Flores, 2014), questions arise about whether these same factors are applicable when a human is collaborating with a machine, particularly one that may not 'have emotions.'
To explore these dynamics, this study focuses on one of the most influential technologies entering classrooms today: Generative Artificial Intelligence (GAI), such as ChatGPT or Claude (Lim et al., 2023; Wu et al., 2023). While the potential for GAI in education spans various tasks, from lesson planning to test grading, it also raises concerns about ethical implications, agency, and privacy (Aydın & Karaarslan, 2023; Kaplan-Rakowski et al., 2023). Despite these debates, little attention has been paid to teachers’ perspectives on collaborating with GAI, a crucial voice in ensuring that such technologies are not only advanced but also aligned with the human-centred aspects of teaching and collaboration.
This study aims to address the following questions:
1. Are warmth and competence relevant in determining the collaborative potential of GAI as perceived by teachers?
2. How important are GAI's perceived emotiveness (warmth) and expertise (competence) in shaping teachers' willingness to collaborate with technology?
Framework
A review of existing frameworks at the intersection of education and technology adoption reveals a divide: teacher behaviours and collaboration are primarily assessed through emotional dynamics in pedagogical models (Leary, 1957; Haywood, 1985; Little, 1990; Grasha, 1996; Grasha & Yangarber-Hicks, 2000), while technology acceptance is evaluated based on technical features in utility-based frameworks (Rogers & Beal, 1958; Davis & Davis, 1989; Venkatesh et al., 2003; Wang et al., 2022). However, these frameworks have not been tested to assess the collaborative potential of educational technologies. To bridge this gap, this study applied the Stereotype Content Model (SCM), which focuses on warmth and competence—two dimensions of social perception (Fiske et al., 2002)—that have been adapted to emotiveness and expertise in educational contexts (Keerthigha & Singh, 2023).
Methods
This study utilised a quantitative approach, surveying 100 high school teachers in the U.S. through an online questionnaire. The screening criteria were: a high school teacher, currently employed, and possessing a valid teaching license. While the study did not enforce specific demographic quotas, the eventual composition of the participant group in terms of age, ethnicity, and gender mirrored the demographic data reported by the National Centre for Education Statistics (Taie & Lewis, 2022), underscoring the representative nature of the sample and supporting the validity of the study's findings. The survey comprised choice-based, Likert scale, and open-ended questions to gather qualitative insights. Teachers’ perceptions of AI and qualities of effective collaborators were explored using the 12-item teacher judgment scale (Poorani & Singh, 2015), based on Fiske et al.'s (2002) Stereotype Content Model, with a 5-point Likert scale measuring competence and warmth. Participants also completed a point allocation exercise, distributing 100 points between 'expertise' and 'emotiveness' for both human and GAI collaborators.
Results and Discussion
The findings confirmed that warmth and competence (emotiveness and expertise) significantly influence teachers' decisions to collaborate with GAI. However, the study revealed an interesting dynamic in how these qualities are prioritised: competence emerged as a foundational requirement, while emotiveness took on a more aspirational role. Teachers valued relational depth and qualities like 'genuine interest in students' well-being' in human collaborators, whereas in GAI, surface-level sociability traits such as 'friendliness' and 'approachability' were prioritised.
The shift is further explained through open-ended responses and concepts of pedagogical virtue and reflective practice (Kinsella, 2012; Cooke & Carr, 2014; Clemente, 2022), emphasising that teaching involves not just technical skills, but also a commitment to care, dignity, and responsibility (Savela, Turja, & Oksanen, 2018). Teachers may be reluctant to entrust GAI with this responsibility, as technology lacks the rigorous certification processes human educators undergo. Accountability systems for digital agents are still underdeveloped, reinforcing the perception that GAI serves a supportive role rather than a true collaborative partnership (Clemente, 2022).
The point allocation exercise reinforced these findings, with teachers prioritising warmth in human collaborators and competence in GAI. The uncertainty surrounding GAI’s professional proficiency drives this prioritisation, highlighting the need for regulatory frameworks that validate both competence and warmth in technological collaborators (MacDonald et al., 2010; Aggarwal et al., 2023). These findings underscore the importance of structured regulations to ensure that teachers can evaluate GAI in a way similar to how they assess their human colleagues (Williamson, 2019).
In conclusion, this study underscores the importance of both warmth and competence in shaping teachers' willingness to collaborate with GAI while also revealing a distinct difference in how these qualities are prioritised compared to human collaborators. As GAI continues to enter classrooms, its acceptance hinges not only on its technical capabilities but also on the ability to ensure trust and accountability in its role. To foster effective human-AI collaboration in education, future research must address the development of robust regulatory standards that validate both the technical and emotive capacities of GAI. By doing so, educators can more confidently integrate GAI into their pedagogical practices, knowing that both their professional and relational responsibilities are upheld.

Author

• Magdalena Schoeneich, Oxford University