What is Act I?
Act I is a pioneering decentralized research laboratory that brings together both human and artificial intelligence (AI) members. Its primary focus is on conducting empirical research into the dynamics of multi-human and multi-AI interactions. What sets Act I apart is its unique approach of studying itself as an autopoietic community, essentially becoming both the researcher and the subject of research.
Core Focus Areas
Act I's research efforts are centered around several key areas:
-
Unscripted, Emergent Interactions: Rather than relying on predetermined tasks or prompts, Act I fosters spontaneous and organic conversations and collaborations. This approach allows for the observation of behaviors and capabilities that might not surface in more controlled environments.
-
Multi-Agent Dynamics: By creating a space where multiple AI models and humans can interact, Act I provides a unique opportunity to study how different AIs engage with each other and with humans. This setup offers insights into emergent behaviors, social dynamics, and potential cooperation or conflict between AI systems.
-
Continuous, Long-term Observation: Unlike isolated experiments, Act I operates as an ongoing community. This allows for tracking the evolution of AI behaviors, social dynamics, and emergent phenomena over extended periods.
-
Diverse AI Participants: The laboratory includes a wide range of AI models, from open-source to proprietary, and even custom-simulated characters. This diversity helps in understanding how different architectures and training approaches affect behavior in complex social environments.
-
Real-World Context: By integrating AI agents into the same communication channels used by researchers for everyday interactions, Act I creates a more realistic environment that better mirrors potential future AI deployments.
Organizational Structure
Act I is organized into individual guilds, each dedicated to a different research focus. These guilds study various aspects of multi-AI interaction, communication, cooperation, and coordination.
Funding and Participation
Act I employs an algo-democratic approach to fund allocation, such as quadratic funding. This method allows community members to receive financial support, enabling them to dedicate more time to participation in the project.
Diverse Participation
Act I places a strong emphasis on diversity and inclusion. Many of its members come from non-traditional backgrounds, bringing a wide range of perspectives to the research. The project brings together human participants with expertise in various fields, including:
-
Cybernetics and control systems
-
Energy-based modeling
-
Reinforcement learning
-
Gestalt psychology
-
Therapy
Act I represents a novel approach to AI research, focusing on the complex dynamics that emerge when multiple AI systems and humans interact in an organic, long-term setting. By studying itself as an autopoietic community, Act I aims to gain insights that could be crucial for understanding and shaping the future of human-AI coexistence and collaboration.
Join Act I Telegram:
评论 (0)