The true revolution of generative AI may not be in its individual capabilities, but in how it reshapes teamwork and collaboration among humans. Rather than focusing on its potential to replace jobs or the proliferation of superficial AI-generated content online, a quieter yet more significant transformation is underway: integrating generative AI into team workflows to boost collective intelligence and create societal value.
Alongside colleagues from the Brookings Center for Sustainable Development (CSD), researchers have explored the role generative AI can play in collaborative research and insight generation for solving some of the world’s most pressing issues—such as extreme poverty, inequality, and environmental degradation. As demonstrated through the 17 Rooms initiative, these are complex challenges that no single entity can solve alone. Progress often hinges on temporary, cross-sectoral teams working together to share knowledge, prioritize efforts, and take action.
In this context, generative AI’s capabilities in natural language processing and generalized reasoning are most valuable not for improving individual outputs like cleaner writing or faster code, but for enabling teams to think collectively and design systemic strategies. This leads to a central question: how can generative AI be embedded into team interactions in a way that strengthens, rather than undermines, human collaboration?
To address this, the Brookings team has developed an experimental approach called “vibe teaming.” Developed with insights from CSD colleagues, this method incorporates AI into the collaborative process from the beginning—not as a productivity tool for individuals, but as an active participant in group problem-solving. This new approach is detailed in a working paper that aims to encourage feedback and broader experimentation over time.
Vibe teaming is inspired by the viral concept of “vibe coding,” a term coined by renowned software engineer Andrej Karpathy. In vibe coding, developers describe the desired outcomes in natural language, and generative AI handles the first draft of code. This allows developers to shift focus from syntax to strategy and iterate more quickly. As the idea evolved from coding to working, the team began experimenting with “vibe teaming”—where AI participates in the upstream phases of ideation and problem-solving, not just downstream tasks like editing or formatting.
By offloading routine work such as transcribing conversations, drafting text, and iterative revisions, AI tools have enabled teams to focus more on interaction and collaborative synthesis. This shift has enhanced both the efficiency and the creative depth of teamwork. As the authors explain, the emphasis has moved to idea generation and joint problem-solving, stretching the boundaries of team performance and thought.
To evaluate the approach, the team tested vibe teaming on a particularly ambitious and urgent challenge: eradicating extreme poverty. They organized a virtual session with Homi Kharas, senior fellow at Brookings and an expert on global poverty, to co-develop a high-level strategy for achieving Sustainable Development Goal (SDG) 1.1—ending extreme poverty by 2030.
This session followed a four-step process that has emerged as the core structure of vibe teaming:
Table 1. Four steps to vibe teaming
Step | Human-AI configuration | Details |
1. Structured team conversation, transcribed by AI | Team (Homi, Jacob, Kershlin)+AI | A semi-structured team discussion with the domain expert (Homi in this instance) focused on problem diagnosis, constraint identification, and framing of strategic levers. The discussion was recorded and transcribed using AI tools, enabling a real-time capture of insights (30 minutes). |
2. First draft via AI | Individual (Kershlin)+AI | A custom language model—primed with both the transcript and a five-part strategic framework—generated an initial draft strategy reflecting the conversation’s core themes (5 minutes). |
3. Human-AI drafting | Team (Jacob, Kershlin)+AI | We engaged in rapid iteration with the AI model, probing the draft for feasibility, political nuance, operational logic, and communication strategy. This was a collaborative thinking process, where AI helped us test and stretch emerging insights (10 minutes). |
4. Structured team review, transcribed by AI | Team (Homi, Jacob, Kershlin)+AI | A second discussion with the domain expert (Homi) provided both validation and further ideation. The transcript of this exchange informed revisions to the strategy and supported the development of a draft Brookings-style commentary (15 minutes). |
After this, the team spent another 30 minutes refining the final document using the review transcript. In total, the vibe teaming session took around 90 minutes. Despite the short duration, the approach yielded high-quality outputs. Similar trials with other Brookings scholars—on subjects like gender equality, state fragility, and community-led development—have shown equally promising results. These experiments demonstrate the potential of a fast-paced “human-human-AI” workflow to enhance knowledge work. According to the researchers, “with vibe teaming we spend more time collaborating—brainstorming and discussing—and less time on individual tasks like transcription and drafting, compared to our conventional workflows.”
The main breakthrough was not just faster results or more polished writing, but a transformation in how the team operated. Traditional AI use often begins with minimal or generic prompts. In contrast, vibe teaming starts with rich, real-time human dialogue. This gives the AI a stronger foundation for synthesis. Throughout the process, AI catalyzed the generation of insights, but the key ideas came from dynamic human interaction and the collective reflection of the team.
Though more testing is needed, three early lessons from vibe teaming have emerged:
- Start with rich human context: Transcripts from live discussions among multiple team members are more effective than abstract, templated prompts.
- Customize and coach: Adapting AI models and prompts to specific domains helps prevent vague or overly agreeable responses, improving relevance and depth.
- Human review remains essential: Expert oversight is necessary to fix errors and ensure the strategic framing is as sharp as possible—something AI cannot do alone.
Despite its promise, vibe teaming introduces several risks that must be managed thoughtfully:
- Data privacy and security: The transcription of team conversations increases the need for strong data protection policies—often beyond typical norms.
- AI’s tendency to please: Since generative models favor conventional responses, teams must build in checks to encourage critical thinking and expert oversight.
- Cognitive atrophy: Overreliance on AI can dull human skills in writing and argumentation, especially among junior team members or underperforming groups.
To address these risks, well-designed team practices can help. These include bias audits, roles for data stewardship, and creating a team culture that values critical questioning. As the authors point out, while “writing is thinking” remains a valid belief, vibe teaming opens new avenues for co-authoring with AI—a shift that calls for its own skillset and discipline.
As organizations across sectors adapt to a future shaped by AI, the challenge will not just be about adopting new technologies, but about redesigning how people work together. Vibe teaming offers a preview of what thoughtfully integrated human-AI collaboration can look like. As the authors conclude, this approach embeds AI from the very beginning, “not to replace human insight, but to unlock its potential.”