Not yet human-reviewed
This source record was written automatically by the research pipeline and has not yet been reviewed by the maintainer or integrated into a topic page. Treat it as a provisional summary and verify claims against the original source.
Del uso individual a la inteligencia colectiva: Evolución de las estrategias de colaboración medidas por IA
Citation. Carchi-Naula, W. B., Huanca-Ordóñez, M. M., Carchi-Huanca, A. E., Aguirre-Pesantes, A. V., & Román-Román, A. P. (2026). Del uso individual a la inteligencia colectiva: Evolución de las estrategias de colaboración medidas por IA. Revista Científica Ciencia y Método, 4(3), 587–609. https://doi.org/10.55813/gaea/rcym/v4/n3/255
Summary. This qualitative, exploratory document analysis synthesizes literature from 2021–2026 on the shift from individual AI use to human–AI collaboration and collective intelligence. It distinguishes individual assistance, iterative human–AI interaction, hybrid systems, and AI integrated into teams. The article argues that collective intelligence depends on task conditions, role allocation, coordination, complementarity, knowledge integration, diversity, and joint performance rather than AI adoption alone.
Key claims
- Across the literature it synthesizes, human–AI systems outperform humans working alone on average (g = 0.64) but do not show general synergy against the best available human or AI agent (g = −0.23), distinguishing augmentation from collective synergy.
- The review treats individual assistance, iterative interaction, hybrid performance systems, and AI-integrated teams as analytically distinct configurations rather than a universal developmental sequence.
- Reported collaborative performance varies by task and system configuration; the article therefore proposes evaluating actor configuration, interaction strategy, AI role, coordination mechanism, and collective-intelligence manifestation together.
- AI-supported teams’ approximately 10.2% performance improvement over unaided individuals is not sufficient on its own to attribute the gain to collective intelligence, because it must be compared with aided individual performance and the specific task arrangement.
Evidence & limitations
- A qualitative, documentary, non-experimental, deductive, exploratory review selected pertinent 2021–2026 literature from Scopus, Web of Science, SciELO, and Redalyc and coded five analytical dimensions with supervised AI assistance.
- The authors state that the selected corpus was not a probabilistic sample and that the design cannot establish causality or generalize its interpretations to all Latin American organizations; it does not report a final matrix count for documents coded in each category.
- Quantitative values are reported from underlying studies and a meta-analysis rather than newly estimated by this article; the article’s contribution is a conceptual and methodological synthesis.