A Digital Laboratory for Social Skills &
Reciprocity Assessment 🔬
An AI-driven research platform designed to measure and develop core social competencies in children. By simulating complex social interdependencies, the system provides a data-rich environment for analyzing reciprocity, social information processing, and adaptive behavior.
The CO-OP World environment creates a non-competitive space where progress depends on mutual support. While the child collects coins and the virtual partner (NPC) collects ice cubes, they soon encounter locked rewards that can only be opened by the other player. This interdependence teaches children that helping others, even at a personal cost of points, is the key to long-term success and mutual benefit.
At the heart of the system is a sophisticated NPC that acts as a social mirror. Using built-in strategies like "Tit-for-Tat," Altruism, or Egoism, the virtual player creates diverse social scenarios to test a child's adaptability. Our latest integration of Large Language Models (LLMs) allows the NPC to explain its decisions in simple, natural language, helping children bridge the gap between action and social understanding.
Every interaction within CO-OP World is captured and analyzed on our GCP-hosted backend. The system tracks key performance indicators such as Positive and Negative Reciprocity, providing researchers with a clear view of a child's social decision-making patterns. Through sophisticated clustering and prediction models, the dashboard transforms raw game data into actionable clinical insights.
Contact sarned@biu.ac.il to get more information about the project