Weekly AI Alignment Paper Reading Groups
AI Alignment Paper Dissection – 10-Week Event
(Source: https://beta.ai-plans.com)
🕒 Time: Fridays, 15:00 UTC (1.5-hour sessions)
📍 Platform: Discord: https://discord.gg/dv9aNAZFSg
Each Reading Group will be led by different people and have their own research focus. Pick the ticket for the Research Area that interests you the most.
If you're interested in leading a reading group, email kabir@ai-plans.com
AI Alignment Through Ethics
📖 Focus Areas: Agent Foundations, Decision Theory, and Ethics (papers from AI-Plans collection)
This is a 10-week recurring event where we’ll critically analyze three papers from each of the above focus areas. At the end of the program, we’ll compile key insights into a literature review paper using LLMs and structured editing.
How It Works:
After each Friday session, a poll with 4–5 paper choices will be posted.
The most voted paper will be finalized by Sunday for the next session, as the one we'll read.
Each participant must read the selected paper and prepare 15 thought, suggested distribution: 7 vulnerabilities, 4 strengths, 3 general thoughts, 1 key insight.
During the session, we’ll discuss these points, with open commenting and debate.
Final Review & Literature Paper
In the final (10th) session, we’ll revisit all discussions with a broader perspective, distill insights into a literature review paper, and explore its potential applications.
You can be a co-author on the paper we'll make, as a literature review. We'll find the most pressing problems in the field and make them available to current and prospective AI Ethics researchers, so they can work on them.
Commitment & Expectations
Expect to dedicate 4–5 hours per week to thoroughly reading the paper.
Sessions are interconnected, so consistent attendance is highly encouraged.
If you're serious about AI Alignment research and willing to engage deeply, this will be an enriching experience! Feel free to reach out if you have any questions. 🚀
MCP & Persistent Memory
Led by Graham dePenros