{"slug": "does-ai-make-teams-better-at-working-together", "title": "Does AI Make Teams Better at Working Together?", "summary": "A synthesis of three studies involving more than 480 learners, co-authored by Rónan Fulton, Alana McCarthy, and Michael Hogan, finds that the effectiveness of AI in team collaboration depends more on interpersonal trust and critical engagement than on the technology's power. The highest-performing groups treat AI output as something to verify, not copy, and trust determines whether AI boosts creativity or fuels conflict. The findings suggest that designing AI for collaborative learning is a psychological challenge as much as a technical one.", "body_md": "######\n[Artificial Intelligence](/us/basics/artificial-intelligence)\n\n# Does AI Make Teams Better at Working Together?\n\n## How a team uses AI matters more than how powerful it is.\n\nPosted August 1, 2026\n[\nReviewed by Margaret Foley\n](/us/docs/editorial-process)\n\n### Key points\n\n- Trust determines whether AI strengthens a team's creativity or fuels conflict within it.\n- The highest-performing groups treat AI output as something to verify, not something to copy.\n- AI helps most when it combines real-time participation with feedback and reflection.\n\n*Co-authored by Rónan Fulton, Alana McCarthy, and Michael Hogan. *\n\nGenerative [artificial intelligence](https://www.psychologytoday.com/us/basics/artificial-intelligence) (GenAI) is rapidly becoming part of collaborative learning. As GenAI tools become increasingly capable, discussion often centres on the technology itself: whether it will improve learning or undermine it. Yet this framing overlooks a more fundamental issue. Consider a surgical team. Access to sophisticated instruments does not determine whether an operation is successful. The outcome depends on how well the team communicates, coordinates, and trusts one another while using those tools. AI-supported collaborative learning is no different: Like a scalpel, AI proves its worth only in skilled, trusting hands.\n\nThree recent empirical studies, involving more than 480 learners across higher [education](https://www.psychologytoday.com/us/basics/education), suggest that the effectiveness of AI depends less on what the technology can do than on how people work with it. Together, they point towards an important conclusion: Designing AI for collaborative learning is not simply a technical challenge, but also a psychological one.\n\n## Interpersonal trust shapes the value of AI-supported collaboration\n\n[Luo and colleagues (2025)](https://link.springer.com/article/10.1007/s10639-025-13527-3) surveyed 308 university students working in 66 collaborative groups to investigate how trust, collective efficacy, and task conflict influenced team [creativity](https://www.psychologytoday.com/us/basics/creativity) when generative AI was introduced. Rather than treating [collaboration](https://www.psychologytoday.com/us/basics/teamwork) as a single outcome, the study examined the social conditions that determine whether AI benefits group work. Trust was measured through students' perceptions of whether group members could openly discuss difficulties, approach tasks professionally, and rely on one another to fulfill their responsibilities. Task conflict reflected disagreements about how work should be organized, and whether AI-generated suggestions or human judgment should guide decisions.\n\nThe findings reveal AI's double-edged influence on collaboration. Students who perceived AI as intelligent developed stronger collective efficacy, a shared [confidence](https://www.psychologytoday.com/us/basics/confidence) in their group's ability to succeed, which in turn promoted team creativity. At the same time, AI also increased task conflict, creating disagreements that reduced creative performance. Trust proved to be the critical difference. Groups with stronger interpersonal trust were better able to turn AI-supported collaboration into creative outcomes, because trust strengthened the link between AI use and collective efficacy. Yet the study also highlights an important limitation. Trust was treated as something groups already possessed rather than something AI actively helped to develop. If trust determines whether AI enhances collaboration, future educational AI may need to support the interpersonal processes that enable groups to work confidently together.\n\n## Critical engagement distinguishes successful collaboration with AI\n\n[Lehtinen and colleagues (2026)](https://onlinelibrary.wiley.com/doi/10.1002/jcal.70256) observed 75 pre-service teachers collaborating with generative AI to design lesson plans. Using video recordings and process mining, the researchers examined how higher- and lower-performing groups interacted with AI during collaborative lesson design.\n\nThe behavioural differences between the groups were striking. Higher-performing groups typically drafted sections of their lesson plans, searched external sources to verify AI-generated information, and then returned to revise their work in light of what they had found. They consulted online resources more than twice as often as lower-performing groups, consistently corroborating AI responses before using them. By contrast, lower-performing groups repeatedly re-prompted ChatGPT when initial responses were unsatisfactory, copied AI-generated text directly into their lesson plans, and then edited the wording afterwards rather than independently verifying the content.\n\nThe difference was not simply one of ability, but of interaction. Successful groups treated AI as something to think with rather than something to think for them. This finding carries an important implication for educational AI design. If critical engagement improves collaborative learning, AI systems should not simply reward efficient prompting by producing increasingly polished responses. Instead, they should encourage learners to compare evidence, justify decisions, and verify information before progressing.\n\n## AI is most effective when it supports the whole collaborative process\n\n[Gyasi and colleagues (2025)](https://www.jstor.org/stable/48827992?seq=1) experimentally compared three approaches to human–AI collaboration in online collaborative learning. One condition introduced AI feedback and feedforward, providing groups with personalised summaries of their performance after each task, alongside guidance before the next activity. These included visualisations of how much of each group's discussion stayed on topic, with written feedback. A second condition embedded an AI chatbot directly within group discussions as an additional participant that contributed ideas, posed questions, offered explanations, and helped keep conversations focused. A third condition that combined both approaches consistently produced the strongest outcomes. Groups supported by both an AI partner and AI feedback demonstrated higher collaborative knowledge building, cognitive engagement, socially shared regulation, and overall group performance than those receiving either intervention independently or no AI support. These findings suggest that collaborative learning benefits when AI supports the entire cycle of planning, monitoring, evaluating, and adapting group work, rather than performing a single isolated function. In short, AI works best when its design reflects that integrated process.\n\nThese broader principles become even clearer when considered alongside [Xu et al.'s (2026) study](https://www.psychologytoday.com/us/blog/in-one-lifespan/202607/does-it-matter-if-your-team-sees-you-using-ai) of shared versus individual AI use in collaborative learning. Whereas the studies discussed here identify trust, critical engagement, and shared regulation as the foundations of successful AI-supported collaboration, Xu and colleagues demonstrate how these qualities are strengthened when AI use is visible to the whole group. When students prompted, evaluated, and revised ChatGPT responses collectively, AI became a shared object for discussion that supported mutual awareness and negotiation. By contrast, when students used AI privately before group meetings, opportunities for collective evaluation and shared [decision-making](https://www.psychologytoday.com/us/basics/decision-making) were reduced. Together, these findings suggest that effective AI-supported collaboration depends not only on what AI supports, but also on how its use is integrated into collaborative activity.\n\n[Intelligence](https://www.psychologytoday.com/us/basics/intelligence)Essential Reads\n\n## Designing AI means designing for collaboration\n\nTaken together, these studies point in the same direction. Trust, critical engagement, and shared regulation do not emerge automatically when groups are given access to AI. They are conditions that need to be actively cultivated through educational practice and, increasingly, through the design of AI itself.\n\nThis represents an important shift in how AI-supported collaborative learning is often discussed. Rather than asking whether AI can improve collaboration, educators and designers may need to ask how AI can support the psychological and social processes that have always underpinned effective collaboration. The evidence reviewed here suggests that the future of AI in education is unlikely to depend solely on building more capable systems. Instead, it may depend on designing AI that helps groups communicate more effectively, think more critically, and regulate their learning together.\n\n*Rónan Fulton and Alana McCarthy are researchers working at the University of Galway.*", "url": "https://wpnews.pro/news/does-ai-make-teams-better-at-working-together", "canonical_source": "https://www.psychologytoday.com/us/blog/in-one-lifespan/202608/does-ai-make-teams-better-at-working-together", "published_at": "2026-08-01 17:34:56+00:00", "updated_at": "2026-08-01 17:35:09.269407+00:00", "lang": "en", "topics": ["artificial-intelligence", "generative-ai", "ai-ethics"], "entities": ["Rónan Fulton", "Alana McCarthy", "Michael Hogan", "Luo", "Lehtinen"], "alternates": {"html": "https://wpnews.pro/news/does-ai-make-teams-better-at-working-together", "markdown": "https://wpnews.pro/news/does-ai-make-teams-better-at-working-together.md", "text": "https://wpnews.pro/news/does-ai-make-teams-better-at-working-together.txt", "jsonld": "https://wpnews.pro/news/does-ai-make-teams-better-at-working-together.jsonld"}}