Google research finds: Chinese AI models replicate human collective intelligence mechanisms

Source: CritpoTendencia Original Title: Chinese AI Models Replicate Human Collective Intelligence, According to Google Original Link:

Implementation of Social Thinking in AI Models

In recent artificial intelligence advancements, a research team at Google has discovered that advanced models developed by DeepSeek and Alibaba Cloud exhibit cognitive patterns comparable to human collective intelligence.

According to the researchers, this finding indicates that internal multi-perspective views and diverse approaches, beyond simple increases in computational power, may be the decisive factors in the capability expansion of contemporary AI models.

Within this framework, a study published on the open repository arXiv analyzes the DeepSeek R1 model and Alibaba Cloud’s QwQ-32B, aiming to understand how they handle complex reasoning.

Experiments show that these AI systems generate internal debates among multiple agents, each with different personality traits and expertise areas. The researchers refer to this behavior as social thinking, emphasizing that the interaction, comparison, and collaboration among these internal agents significantly extend the model’s ability to handle complex problems.

Diversity and Progress of Chinese Open-Source Models

According to the authors, this simulated debate mechanism establishes computational parallelism with human collective intelligence, where structured diversity improves decision-making and problem-solving. This phenomenon suggests that AI is beginning to replicate the dynamics of human communities in an emergent manner.

The report highlights that the diversity of internal perspectives within the models is a primary driver of their intelligence and effectiveness in complex tasks, thereby relativizing the exclusive role of increased computational power.

Similarly, the research underscores the growing relevance of open-source models developed in China, such as DeepSeek and QwQ-32B. In particular, these systems have become key tools in cutting-edge interdisciplinary research in the United States, reflecting strengthened scientific collaboration in the global AI field. Against this backdrop, Alibaba Cloud, the cloud computing division of Alibaba Group Holding, has reinforced Asia’s position in disruptive technology development, while the international scientific community recognizes these models’ contributions to AI progress.

A New Paradigm in AI Development

Overall, the Google team concludes that collective intelligence is not exclusive to humans. When AI models incorporate internal debates and cognitive diversity mechanisms, they begin to replicate collaborative and adaptive psychological processes.

Although this research has not yet undergone peer review, its results represent an important advancement in understanding how AI can transcend traditional limitations and expand its influence in science, technology, and complex decision-making fields.

Based on these findings, the development of models like DeepSeek and Alibaba QwQ-32B indicates that the next frontier of artificial intelligence will not be solely defined by computational capacity but by the structured integration of multiple perspectives and capabilities, simulating the collective operation of human thought.