As generative artificial intelligence becomes more embedded in workplace practices, questions arise about its deeper implications for collaboration, performance, and professional identity. In their groundbreaking study “The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise”, Fabrizio Dell’Acqua, Charles Ayoubi, Hila Lifshitz, Raffaella Sadun, Ethan Mollick, Lilach Mollick, and colleagues explore these questions through a large-scale field experiment at Procter & Gamble. This research offers rich insights into how AI affects not just productivity, but also the social and cognitive dynamics that underpin effective teamwork.
The authors conducted a pre-registered experiment with 776 professionals at Procter & Gamble, randomly assigning participants to one of four work configurations: individuals working alone, teams of two humans, individuals with AI, and teams of two humans with AI. All participants engaged in real-world product development tasks using P&G’s existing innovation processes. The study’s central hypothesis was that AI could do more than automate tasks—it could serve as a collaborative partner, replicating or even enhancing the benefits typically associated with human teamwork.
One of the key findings is that individuals working with AI produced results of a similar quality to those created by two-person teams without AI. This suggests that AI can substitute for some of the functional advantages of human collaboration, such as combining diverse expertise and generating high-quality output. Moreover, participants using AI spent significantly less time on their tasks, indicating efficiency gains without compromising quality. These results reinforce the view that generative AI, when integrated into cognitive workflows, can amplify individual capabilities in a way that mirrors collaborative synergy.
A particularly striking aspect of the study is the way AI impacts professional silos. In typical settings, R&D specialists tend to propose more technical solutions, while commercial professionals focus on market-oriented proposals. The presence of AI disrupted this pattern. When AI was involved, participants—regardless of their professional background—produced more balanced solutions that integrated both technical and commercial considerations. This finding suggests that AI can act as a boundary-spanning agent, enabling employees to think beyond their core domain and engage in more interdisciplinary problem-solving.
In addition to performance and expertise sharing, the authors also studied the emotional effects of working with AI. Participants who collaborated with AI reported higher levels of positive emotions such as enthusiasm and energy, and lower levels of negative emotions such as anxiety and frustration. In fact, individuals working with AI showed emotional responses that matched or exceeded those of human teams without AI. These findings challenge the assumption that AI erodes the social dimensions of work. On the contrary, the language-based nature of generative AI tools appears to create a more engaging and emotionally supportive environment for users.
The study also examined the likelihood of exceptional performance. Teams that combined human collaboration with AI support were significantly more likely to produce solutions ranked in the top 10% of all submissions. This suggests that the combination of AI augmentation and human collaboration may be especially powerful in generating breakthrough ideas. While individuals with AI performed strongly, the synergistic effect of teaming up with both humans and machines seems to offer unique advantages for complex problem-solving.
Despite these positive outcomes, the researchers emphasize the importance of thoughtful integration. Participants in the study were relatively inexperienced in AI prompting, and the tools used were not optimized for team-based workflows. As organizations adopt more sophisticated tools and users become more skilled, the benefits could increase. However, they also caution that AI’s standardizing effects could reduce the diversity of ideas if not carefully managed. The emotional and cognitive dependencies AI creates must also be considered when designing future work environments.
Ultimately, the study reframes our understanding of AI’s role in knowledge work. Rather than viewing generative AI as just another tool, the authors argue that it should be understood as a “cybernetic teammate”—a feedback-driven partner capable of influencing cognition, expertise sharing, and emotional engagement. This has far-reaching implications for how organizations structure teams, train employees, and assess performance in an AI-enhanced workplace.
The full study “The Cybernetic Teammate: A Field Experiment on Generative AI Reshaping Teamwork and Expertise” by Dell’Acqua et al. is available online at SSRN.