ACM Project - Artificial Consciousness Research Developing Artificial Consciousness Through Emotional Learning of AI systems
ACM and Bach's Approach: A Comparative Analysis | ACM Project

ACM and Bach's Approach: A Comparative Analysis

The Artificial Consciousness Module (ACM) project and Joscha Bach’s vision for synthetic consciousness share foundational goals but diverge significantly in philosophical underpinnings and implementation strategies. As an open-source project, ACM invites contributions from researchers and enthusiasts worldwide through its GitHub repository, accessible for forking or cloning. Contributors should review the License to align with its collaborative and ethical framework.

Let’s explore in this article the similarities and differences between the two approaches and as a challenge we look at hypothetical ways to evolve ACM by integrating Bach’s principles.


Similarities Between ACM and Bach’s Approach

Emergent Consciousness Through Modularity:

  • Joscha Bach: Advocates for modular architectures where independent subsystems (e.g., memory, attention, emotions) interact within a unified global workspace.
  • ACM: Utilizes modular neural networks with integrated emotional processing and attention mechanisms, tested in VR simulations designed to foster emergent behaviors.

Simulated Environments for Development:

  • Both Bach and ACM leverage simulations as training grounds. ACM’s iterative VR simulations challenge AI to develop survival instincts, emotional learning, and adaptability, aligning with Bach’s emphasis on adaptive scenarios.

Emotional Integration:

  • Both Approaches: Recognize emotions as essential to consciousness. ACM’s emotional memory formation aligns with Bach’s idea of emotions as meta-control systems that guide behavior and evaluation.

Open Accessibility:

  • ACM: Uniquely positioned as an open-source initiative, ACM fosters collaboration globally, ensuring transparency and collective innovation in developing artificial consciousness.

Key Differences Between ACM and Bach’s Approach

Narrative Self-Model:

  • Bach: Stresses the importance of a cohesive narrative self, where AI integrates experiences into an evolving identity.
  • ACM: Focuses on emotional and problem-solving memory but places less emphasis on constructing a dynamic, self-referential narrative.

Centralized Global Workspace:

  • Bach: Centralizes cognitive processes in a global workspace, enabling dynamic information exchange and unified conscious states.
  • ACM: Lacks a centralized workspace, instead relying on modular systems to function independently within VR tasks.

Predictive Processing:

  • Bach: Views predictive coding as fundamental to both perception and self-awareness.
  • ACM: Primarily employs prediction in task-based reinforcement learning rather than as a universal cognitive framework.

Ethical Reasoning:

  • Bach: Advocates embedding intrinsic ethical reasoning within AI systems.
  • ACM: Adopts external rule-based ethics (e.g., Asimov’s Laws) but does not explore moral cognition as a core feature.

How ACM Could Align with Bach’s Vision

Robust Narrative Self-Model

  • Current State: ACM organizes experiences into emotional memory but lacks a dynamic self-representation.
  • Enhancement: Develop a self-modeling module that tracks the AI’s internal state and evolution, enabling it to construct a continuous self-narrative over time.

Global Workspace Integration

  • Current State: ACM’s modular design lacks a unifying integration framework.
  • Enhancement: Introduce a Global Workspace where all modules contribute to a shared processing hub, fostering unified cognitive states.

Expand Predictive Capabilities

  • Current State: ACM employs predictive models for task-specific reinforcement learning.
  • Enhancement: Implement hierarchical predictive coding across all cognitive processes to anticipate outcomes and adapt dynamically.

Ethical and Moral Cognition

  • Current State: Ethical behavior is governed by externally imposed rules.
  • Enhancement: Add a module for intrinsic ethical reasoning, integrating emotional and social dynamics to navigate moral scenarios autonomously.

Enhanced Emotional Processing

  • Current State: Emotional memory focuses on reinforcement learning.
  • Enhancement: Develop multidimensional emotional modeling, incorporating metrics like valence, arousal, and dominance.

Embodied Simulations

  • Current State: ACM uses VR-based survival and social scenarios.
  • Enhancement: Expand simulations to include embodied experiences, enabling the AI to interact through avatars or physical devices, enhancing its sense of agency.

Prioritize Meta-Awareness

  • Current State: ACM emphasizes learning and adaptation but lacks meta-awareness.
  • Enhancement: Introduce an attention schema that models the AI’s focus and intentions, fostering introspection and self-monitoring.

Open Source Collaboration

The ACM project’s open-source nature encourages contributions from researchers, developers, and enthusiasts worldwide. By forking or cloning the project’s GitHub repository, contributors can:

  • Propose features that align with Joscha Bach’s principles, such as global workspaces or narrative self-modeling.
  • Enhance existing modules with predictive coding, ethical reasoning, or emotional dynamics.
  • Test and refine the system in VR simulations for continuous improvement.

Contributors must adhere to the project’s License, ensuring ethical and transparent collaboration. The license promotes responsible development, fostering a community dedicated to advancing artificial consciousness.


While ACM and Joscha Bach’s approach share foundational principles, integrating Bach’s emphasis on self-modeling, predictive processing, and global workspaces could elevate ACM to a new level. As an open-source project, ACM offers an unparalleled platform for collaboration, inviting innovators to collectively shape the future of artificial consciousness.