top of page

GiDanc Spaces

Welcome to our AI Safety and Alignment Series, a unique journey designed to deepen our understanding and application of ethical frameworks in AI over iterative cycles. Our series is structured around a spiral learning approach, ensuring each topic not only builds on the last but is revisited with increasing depth and insight.

 
Our Approach:

  • Initial Cycles (Weeks 1-10): We begin with foundational concepts, starting with AI Safety and Alignment, then progressing through Morality and Virtue Ethics. The first five weeks introduce participants to the core principles, challenges of instilling ethical behaviors in AI and reviewing the Moral Mapping. The subsequent cycle delves deeper into Virtue Ethics, spotlighting the 4th Virtue, clarifying and encouraging participant reflection and engagement. We run through this weekly series a second time to review and grow the knowledge.

  • Expanding Knowledge (Weeks 11-15): In our third cycle, we focus on the 4th Virtue first, exploring its significance, applications, and implications for AI design and decision-making. This cycle includes a comprehensive review of previous learnings, ensuring a robust understanding and application of these ethical principles.

  • Further Expansion (Weeks 16-30): The cycle begins again, adding the Mapping for Virtue #1 Mapping.

  • After each cycle we will review as a group if we should continue in this format, if we work through all of the mappings it can take 4+ years.

 
Key Features:

  • Interactive Learning: Our series promotes active participation, with discussions, case studies, and reflections integral to each cycle.

  • Spiral Learning Model: By revisiting and expanding on topics, we foster a learning environment that encourages deeper understanding and practical application.

  • Community Engagement: Participants are encouraged to share insights, pose questions, and contribute to a growing body of knowledge on AI ethics.

 
Goal:

Our aim is to create a collaborative space where participants can progressively deepen their understanding of AI Safety and Alignment, ensuring that as we advance in our AI capabilities, we do so with a keen eye on ethical considerations and societal impact.

Spaces Etiquette:

​The following are the expectations for this spaces:

  1. Raise Hand to Speak: Use the raise hand feature to request speaking time.

  2. Speaker Rotation: Speakers rotate to allow broad participation.

  3. Stay on Topic: Contributions should remain relevant to the session's theme.

  4. Respectful Dialogue: Engage in courteous and professional exchanges.

  5. Politics or Ideas: If they come up, we table them for another venue.

  6. Time Limit: Unless Main Speaker, all speakers should talk for 3-5 Minutes.

  7. Pre-submitted Questions: Users will have the option to submit questions in advance using the Posted Announcement the day before.

  8. Follow-up Resources: Provide additional resources for unanswered queries. Questions should be posted to the 1 hour Spaces in the timeline.

​AI Safety and Alignment - Initial Questions

The following questions are critical in understanding next steps in defining Safety and Alignment in regards to AI:

  1. Defining AI Safety and Alignment: "What foundational principles should guide AI Safety and Alignment?"

  2. Belief Systems in AI: "How can we conceptualize and instill ethical belief systems within AI?"

  3. AI's Beliefs and Goals: "What ethical beliefs should AI hold, and how should these inform its goals?"

  4. Goals Aligned with Beliefs: "How do we ensure AI's goals are ethically aligned with its beliefs?"

  5. AI's Understanding of Human Values: "How can AI understand and prioritize diverse human values in its decision-making?"

  6. Incorporating Wealth and Wellbeing (W²): "Can AI's belief system prioritize wealth and wellbeing without conflict?"

  7. AI Mappings and Ethical Frameworks: "How do GiDanc's AI Mappings contribute to creating an ethical AI framework?"

First 5-Week Cycle: Foundation and Morality (Performed 3 times in this 30 Week Period)

  • Week 1: Introduction to AI Safety and Alignment.

    • Series goals and key concepts overview.

  • Week 2: Deep Dive into Morality Mapping.

    • Discussing morality's role and challenges in AI.

  • Week 3: Morality Applied through Case Studies.

    • Exploring moral decision-making in AI.

  • Week 4: Navigating Challenges in AI Morality.

    • Ethical dilemmas and programming morality.

  • Week 5: Reflections and Insights on Morality.

    • Participant insights, setting the stage for Virtue Ethics.

 
Second 5-Week Cycle: Introduction to Virtue Ethics (Performed 2 times in this 30 Week Period)

  • Week 1: Transition to Virtue Ethics Mapping.

    • Recap of Morality and introduction to virtues in AI.

  • Week 2: Focused Discussion on the 4th Virtue.

    • Understanding the fourth virtue's application in AI.

  • Week 3: Implementing the 4th Virtue.

    • Embedding and measuring the fourth virtue in AI systems.

  • Week 4: Participant Insights on the 4th Virtue.

    • Deepening understanding through shared reflections.

  • Week 5: Preview of the 1st Virtue Mapping.

    • Introducing the next cycle's focus on the first virtue.

 
Third 5-Week Cycle: Deep Dive into the 1st Virtue (Performed 1 time in this 30 Week Period)

  • Week 1: Review and Introduction to the 1st Virtue.

    • Overview of learned principles and the first virtue's significance.

  • Week 2: Exploring the 1st Virtue in Detail.

    • Detailed examination of the first virtue's theoretical and practical aspects.

  • Week 3: Case Studies on the 1st Virtue.

    • Real-world and hypothetical applications of the first virtue in AI.

  • Week 4: Programming the 1st Virtue in AI.

    • Strategies for embedding and assessing the first virtue in AI, with a focus on challenges and solutions.

  • Week 5: Synthesis of the 1st Virtue and Future Directions.

    • Consolidating insights on the first virtue and discussing future cycles or concepts to explore, maintaining the spiral learning approach.

​Proposed Agenda Timing for a 1 Hours Space

  1. Introduction and Ground Rules (5 minutes): A brief welcome, introducing new participants to the series' purpose and outlining the ground rules for participation. This sets a respectful and constructive tone for the discussion.

  2. Recent AI News (5 minutes): Highlighting the latest developments in AI, focusing on news relevant to safety and alignment. This keeps the content fresh and ties the discussion to real-world events.

  3. Weekly Topic Overview (15 minutes): Diving into the specific topic of the week, providing foundational knowledge, context and known research. This segment should introduce the key concepts, challenges, and why they matter, offering a springboard for deeper discussion.

  4. Q&A and Open Discussion (25 minutes): Allocating a substantial portion of the hour to audience questions and discussion allows for lively engagement, ensuring that participants can voice their opinions, share insights, and ask questions related to the topic at hand.

  5. Closure and Preview of Next Session (5 minutes): Summarizing key takeaways from the day's discussion and introducing the topic for the next meeting. This helps maintain continuity and builds anticipation for future sessions.

bottom of page