Skip to content

Stampy / AISafety.info

📋Page Status
Page Type:ResponseStyle Guide →Intervention/response page
Quality:45 (Adequate)⚠️
Importance:50 (Useful)
Last edited:2026-02-02 (4 days ago)
Words:1.3k
Structure:
📊 11📈 0🔗 6📚 1326%Score: 13/15
LLM Summary:AISafety.info is a volunteer-maintained wiki with 280+ answers on AI existential risk, complemented by Stampy, an LLM chatbot searching 10K-100K alignment documents via RAG. Features include a Discord bot bridging YouTube comments, PageRank-style karma voting for answer quality control, and the Distillation Fellowship program for content creation. Founded by Rob Miles as a 501(c)(3) nonprofit.
Issues (2):
  • QualityRated 45 but structure suggests 87 (underrated by 42 points)
  • Links3 links could use <R> components
DimensionAssessmentEvidence
Content CoverageSubstantial280+ live answers, hundreds of drafts
Data SourcesComprehensive10K-100K documents from alignment literature
AccessibilityHighFree web interface, Discord bot, chatbot
Community IntegrationStrongYouTube bridging, karma voting, write-a-thons
Open SourceYes10 public GitHub repositories
MaintenanceActiveGlobal volunteer team + paid editor fellowships
AttributeDetails
NameAISafety.info (also known as Stampy)
OrganizationAshgro Inc (501(c)(3) nonprofit)
FounderRob Miles
Websiteaisafety.info
GitHubgithub.com/StampyAI (10 repositories)
DatasetHuggingFace: alignment-research-dataset
DiscordRob Miles AI Discord (active community)
LicenseMIT (open source)

AISafety.info is a collaborative Q&A wiki focused on existential risk from artificial intelligence, founded by AI safety educator Rob Miles. The project combines human-written educational content with an LLM-powered chatbot, a Discord bot bridging YouTube and Discord communities, and structured programs for content creation.

The site’s core thesis is that “smarter-than-human AI may come soon” and “it could lead to human extinction.” Rather than simply asserting these claims, the wiki provides structured explanations, addresses common objections, and offers pathways for further engagement.

ComponentPurposeTechnology
Q&A WikiHuman-written answers to AI safety questionsWeb frontend (Remix/Cloudflare)
Stampy ChatbotLLM-powered answers with citationsRAG pipeline + GPT models
Discord BotYouTube integration, community moderationPython, modular architecture
Alignment Research DatasetCurated corpus for chatbotHuggingFace, 10K-100K documents
MetricValue
Live Answers280+
Draft AnswersHundreds in development
Content UpdatesOngoing community contributions
Feedback SystemGoogle Docs integration for comments

The chatbot draws from a curated corpus hosted on HuggingFace:

MetricValue
Document Count10,000 - 100,000
Monthly Downloads≈1,600
LicenseMIT
LanguageEnglish

Sources include:

  • Academic: arXiv papers, Arbital
  • Forums: Alignment Forum, LessWrong, EA Forum
  • Organizational blogs: MIRI, DeepMind, OpenAI
  • Individual blogs: Eliezer Yudkowsky, Gwern Branwen
  • Educational: AGI Safety Fundamentals course
  • Video: YouTube playlists on AI safety

The chatbot uses Retrieval-Augmented Generation (RAG) with a three-step process:

  1. Retrieval: Search the alignment-research-dataset for semantically similar chunks using vector embeddings
  2. Context Assembly: Feed relevant text snippets into an LLM’s context window
  3. Generation: Produce a summary with citations to source documents

Dual Response Strategy: Stampy prioritizes human-written answers from the wiki when available, falling back to AI-generated responses for novel questions. This reduces hallucination risk for common questions while maintaining coverage for the “long tail.”

Acknowledged Limitations: The documentation explicitly warns that “like all LLM-based chatbots, it will sometimes hallucinate.” Source citations allow users to verify accuracy.

The Discord bot (StampyAI/stampy) has evolved significantly from its original purpose:

Module System: Rob Miles implemented a “bidding” architecture where different modules compete to handle messages, minimizing computation by only activating relevant handlers.

Key Modules:

  • Question management (Questions, QuestionSetter)
  • Factoid database
  • Wolfram Alpha integration
  • LLM response generation (GPT-4 whitelist available)
  • Alignment Forum search

A distinctive feature is bidirectional integration with Rob Miles’ YouTube channel:

  1. YouTube → Discord: Interesting comments from YouTube videos are posted to Discord, sparking community discussions
  2. Discord → YouTube: Quality responses can be posted back as official YouTube replies

Quality Control via Stamps: The system uses a “stamp” emoji reaction for karma voting. When responses receive enough stamps, they can be posted to YouTube. Critically, stamp value varies by user reputation using a PageRank-style algorithm—users with more stamps have more voting power.

FeatureDescription
Stamp ReactionsKarma voting for response quality
PageRank WeightingVote weight proportional to voter’s reputation
Threshold PostingResponses posted to YouTube when stamp threshold met
Bot IdentityPrevents random users from posting as official channel

Stampy maintains 10 public repositories:

RepositoryStarsPurpose
stampy-ui41Web frontend (TypeScript)
stampy40Discord bot (Python)
stampy-chat15Conversational chatbot (TypeScript)
alignment-research-dataset13Data scraping pipeline (Python)
stampede-Elixir chatbot framework (alpha)
stampy-nlp-NLP microservices (Python)
stampy-extension-Browser extension
GDocsRelatedThings-Google Docs integration
AISafety.com1Issue tracker (54 open issues)
StampyAIAssets-Logos and branding
RoleDescription
FounderRob Miles (YouTube creator, AI safety educator)
EditorsPaid staff from Distillation Fellowship programs
DevelopersVolunteer contributors
CommunityDiscord members, write-a-thon participants

A structured 3-month paid program for content creation:

  • Completed: Two fellowship cohorts
  • Purpose: Train editors to distill complex AI safety content into accessible answers
  • Output: Significant portion of the 280+ live answers
  • Future: Additional cohorts planned pending funding

Community events for collaborative content creation:

  • Format: Multi-day focused writing sprints
  • Example: October 6-9 write-a-thon (third event)
  • Output: Batch content creation and answer improvement

AISafety.info serves as an accessible entry point for people encountering AI risk arguments for the first time:

  • Start with basic questions and progress to advanced topics
  • Find responses to specific objections
  • Understand reasoning behind AI safety concerns
  • Access cited sources for deeper reading

The platform supports AI safety communication:

  • Reference answers when addressing common questions
  • Link skeptics to well-structured objection responses
  • Consistent explanations across audiences
  • Google Docs integration for collaborative editing

While primarily aimed at broader audiences:

  • Entry points into technical literature via dataset
  • Career guidance for field entry
  • Community connections via Discord
StrengthEvidence
Accessible explanationsContent written for general audiences
Quality controlPageRank-style voting prevents low-quality YouTube responses
Community integrationYouTube bridging creates feedback loop
Structured programsDistillation Fellowship produces consistent content
Comprehensive dataset10K-100K documents from major alignment sources
Open sourceAll code publicly available, MIT licensed
LimitationImpact
Chatbot accuracyLLM hallucination risk; users must verify sources
Volunteer capacityDevelopment and content dependent on contributor availability
Opinionated framingPresents AI risk case rather than neutral overview
Dataset maintenanceOngoing work to clean and update sources
Single community perspectivePrimarily reflects EA/rationalist community views
SourceType
Individual DonationsVia website and Every.org
EA CommunityGrants and donations
ManifundProject funding platform
Volunteer LaborPrimary development resource
  • Distillation Fellowship funding for continued content creation
  • Developer time for frontend redesign and chatbot improvements
  • Dataset curation for ongoing maintenance