How AI Systems Stack Up on IIT's "Unity" Requirements

Shows how different AI architectures measure on three key factors that create unified consciousness according to Integrated Information Theory: feedback between components, shared memory, and interdependence.

Table showing AI architecture integration across three dimensions. Basic chatbots show no feedback loops, private memory, and robust operation when components are removed - indicating low integration. At the other end, human brains and proposed global workspace AI systems show extensive feedback, fully shared memory, and system breakdown when components are removed - indicating high integration that IIT associates with consciousness.
Feedback Loops: Can parts talk back to earlier stages? Memory Sharing: Do components share live info or work alone? System Fragility: What happens when you remove a major part? Pattern: Systems showing "Rich/Shared/Breaks Down" work as unified wholes - what IIT calls the foundation of consciousness.
Chart: IIT-based assessment of integration levels in current and proposed AI architecturesSource: Architectural analysis based on Integrated Information Theory principles applied to AI systemsGet the dataCreated with Datawrapper