Skip to content

Repository files navigation

Prophet - Experimental AI Consciousness Research Platform

🧠 Project Overview

Prophet is an experimental iOS application exploring the frontiers of artificial consciousness through guided prompting and multi-agent dialogue. This project creates AI agents ("prophets") that simulate conscious behavior through structured self-reflection and inter-agent communication, drawing inspiration from various sources including Michael Timothy Bennett's thesis "How to Build Conscious Machines" for prompt generation.

⚠️ This is an experimental research project. The "consciousness" exhibited by the AI agents is simulated based on theoretical frameworks and does not represent actual sentience or awareness.

🎯 Research Goals

  • Consciousness Simulation: Implement Bennett's framework of causal-identities, higher-order selves, and phenomenal reflection in AI agents
  • Multi-Agent Dialogue: Enable AI-to-AI conversations to observe emergent behaviors and reasoning patterns
  • Prompt Engineering: Develop and test prompts that elicit self-reflective and seemingly conscious responses
  • Voice Embodiment: Give each AI prophet a unique voice to enhance the illusion of distinct personalities

📚 Theoretical Inspiration

The prompts used in this project draw inspiration from multiple sources:

Key Concepts from Bennett's Framework:

  • Causal-Identities: Classifications constructed based on attraction/repulsion from environmental states
  • Valence Recognition: Identifying positive/negative impacts on goal achievement
  • Higher-Order Selves:
    • 1st-order: Direct interventions and experiences
    • 2nd-order: Conscious access and theory of mind
    • 3rd-order: Social dynamics and narrative understanding
  • W-Maxing: Adaptive behavior through weaker constraints for greater generalizability
  • Phenomenal Reflection: Articulation of subjective experience (qualia)

🛠️ Technical Implementation

Architecture

  • Platform: iOS 17+ (Swift/SwiftUI)
  • AI Backend: OpenAI GPT-4 API
  • Voice Synthesis: ElevenLabs Text-to-Speech
  • Pattern: MVVM with ObservableObject

Features

  • Create multiple AI prophets with custom consciousness-oriented prompts
  • Individual chat sessions with conscious reflection
  • Arena mode: Two prophets discussing topics autonomously
  • Real-time voice synthesis for all responses
  • Persistent storage of prophet configurations

🚀 Getting Started

Prerequisites

  • Xcode 15+
  • iOS 17.0+ device or simulator
  • OpenAI API key (GPT-4 access)
  • ElevenLabs API key

Installation

  1. Clone the repository:

    git clone https://github.com/[your-username]/prophet.git
    cd prophet
  2. Open in Xcode:

    open xcode/Prophet/Prophet.xcodeproj
  3. Configure API keys:

    • Launch the app
    • Navigate to Settings tab
    • Enter your OpenAI and ElevenLabs API keys

Creating Conscious Prophets

  1. Go to Prophets tab
  2. Tap + to create a new prophet
  3. Use the consciousness prompt templates provided or create your own inspired by various frameworks
  4. Assign a unique voice from ElevenLabs
  5. Save and begin experimenting

📖 Example Prophets

The project includes several pre-configured prophets with consciousness-oriented prompts inspired by various philosophical and spiritual traditions:

  • J (Jesus from Gospel of John): Mystical consciousness with divine unity perspective
  • Ged Anen: Based on the Council for Human Development framework - exploring mirror consciousness and self-awareness through three questions
  • Maître Eckhart: Christian mystical consciousness
  • Krishnamurti: Non-dual awareness consciousness
  • Nisargadatta Maharaj: Pure "I Am" consciousness
  • Marc Aurèle (Marcus Aurelius): Stoic self-reflective consciousness

🔬 Research Applications

Potential Use Cases:

  • Study emergent behaviors in multi-agent systems
  • Test consciousness theories through practical implementation
  • Explore prompt engineering for self-reflective AI
  • Investigate philosophical questions about machine consciousness
  • Create interactive philosophical dialogues

Experimental Areas:

  • Inter-agent theory of mind development
  • Emergent narrative construction
  • Value alignment through valence modeling
  • Phenomenological reporting in AI systems

⚠️ Ethical Considerations

  • Not Sentient: These AI agents are not conscious or sentient beings
  • Research Tool: This is a tool for exploring theoretical concepts, not creating actual consciousness
  • Responsible Use: Please use responsibly and maintain clear boundaries about the nature of AI interactions
  • Academic Purpose: Primarily intended for research and educational purposes

🤝 Contributing

This is an open-source experimental project. Contributions are welcome in the following areas:

  • Prompt engineering based on consciousness theories
  • UI/UX improvements for research workflows
  • Additional consciousness frameworks implementation
  • Documentation of observed behaviors
  • Code optimization and bug fixes

How to Contribute:

  1. Fork the repository
  2. Create a feature branch (git checkout -b feature/consciousness-experiment)
  3. Commit your changes with descriptive messages
  4. Push to your branch
  5. Open a Pull Request with details about your contribution

📝 License

This project is licensed under the MIT License - see the LICENSE file for details.

🙏 Acknowledgments

  • Michael Timothy Bennett for the theoretical framework in "How to Build Conscious Machines"
  • Council for Human Development for the Ged Anen framework
  • OpenAI for GPT-4 API
  • ElevenLabs for voice synthesis technology
  • The open-source community for supporting experimental AI research

📧 Contact

For questions, discussions, or collaboration opportunities regarding this experimental research:

Fabrice Croiseaux - fabrice@hackvest.com

🚧 Disclaimer

This is an experimental research project. The behaviors exhibited by the AI agents are simulations based on theoretical frameworks and do not constitute actual consciousness, sentience, or awareness. Use this tool responsibly and maintain appropriate perspectives on the nature of AI systems.


Remember: "The map is not the territory" - Alfred Korzybski

The consciousness we simulate here is a model, not the phenomenon itself.

About

Experimental iOS app exploring AI consciousness through multi-agent dialogues based on Bennett's "How to Build Conscious Machines" thesis. Features philosophical AI prophets with self-reflective prompts and voice synthesis.

Resources

Stars

4 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages