Is a "Phi-Zero" AI architecture viable for safe robotics? (github.com)

A new AI architecture called "Phi-Zero" has been proposed, inspired by the structures of slime molds and siphonophores, aiming to establish safer robotics systems. This design is explicitly non-conscious, relying on three architectural principles: strict feedforward flow without recurrent loops, modular fragmentation that avoids global workspace or lateral communication, and stateless operation that entails a complete memory reset at each time step. By deliberately constructing a system devoid of consciousness, the creators intend to enhance safety and reliability in robotic applications. The significance of Phi-Zero lies in its potential to mitigate risks associated with advanced AI systems, particularly in situations where unpredictability could lead to harmful outcomes. As the AI/ML community grapples with ethical considerations and the consequences of deploying autonomous systems, Phi-Zero presents an innovative approach to ensure safety while still achieving complex tasks. Its technical design could pave the way for more robust robotic solutions, enabling developers to harness AI's capabilities without the existential risks posed by conscious-like systems.