This page proposes additions to the AI Safety Groups Resource Center, drawn from a full read of the ais-groups.slack.com workspace on 2026-09-15: every message in #organizing-discussion (2023-08-17 through 2026-09-14), plus #announcements, #general, #opportunities, #technical-discussion, #governance-discussion, #introductions, #oasis-4, #oasis-7, and the three #pathfinder-f26 channels, along with the substantive threads inside them.

Nothing in the Resource Center or any of its existing subpages was changed. Everything below is a proposal for a human to review and apply by hand.

Counts: 58 new entries, 7 updates to existing entries, 14 already covered. Sections follow the Resource Center's own headings, so each block can be pasted across.

Getting started

• Basic Uni Groups Strategy (doc). Matt Reardon's strategy doc for EA university groups, shared as a useful analogue rather than an AI safety document. Raised by Jesse Gilbert, Slack.
• Group seeding referral form (form). Where to send names of potential founders at universities that have no AI safety group yet, run by Jacob Brinton, Jason Chin, and Thomas Rodskog. Raised by Jacob Brinton, Slack. Belongs next to "Finding Co-Organizers".
• Group naming and acronyms. A long, practical thread on picking a group name, including the UK point that "society" reads more naturally than "club". Raised by Gaurav (CAISH), Slack.
• Fiscal sponsorship through Hack Club (HCB). Pathfinder groups are now introduced to Hack Club as fiscal sponsor at kickoff, which matters for any group whose university will not let it hold external funds. Raised by Jeremy Kintana, Slack. Fellows in Fall 2026 reported real lag between budget approval and money landing in the HCB account, so groups should plan for a gap, Slack.
• Pathfinder budget drafting guidelines (doc) and example budget (sheet). The two documents organizers actually need when writing a funding request. Raised by Jeremy Kintana, Slack.
• Leo Gao's AI safety microgrants (doc). Low-bureaucracy small grants for ambitious AGI-related projects, circulated repeatedly in this Slack. Raised by Gabriel Wu, Slack.
• grantmaking.ai (site). Fast-turnaround $5k to $50k grants for AI x-risk work. Raised by Ryan Kidd, Slack.
• Free bulk books for groups. Kairos has twice run bulk orders that groups can tap for tabling and handouts: Ben Todd's 80,000 Hours book, Slack, and If Anyone Builds It, Everyone Dies through MIRI, Slack. Agus explicitly told groups to have a low bar and ask for up to 100 to 200 copies, Slack.
• Privacy policy templates for small orgs (EA Forum post with templates). Plain-English guide plus GDPR-compliant privacy notice, T&Cs, and volunteer NDA, aimed at small EA and AI safety orgs. Raised by Gergő Gáspár, Slack.

Strategy

• Dillon Nguyen's case against the wide-net fellowship (Slack). A long, careful argument that a wide-net intro fellowship wrecks mid-funnel quality, and that heavy selection plus a small curated cohort gets you the same people better activated. The thread is the most substantive strategy disagreement in the workspace this year and both sides are worth carrying over, so see the Updates section below rather than treating this as settled.
• A barbell strategy for university AI safety groups (doc). Nick Marsh's OAISI memo arguing for investing at the two ends of the funnel and little in the middle, which he thinks has aged well. Raised by Nick Marsh, Slack.
• Georgia Tech AISI retrospective, AY 2025-2026 (LessWrong). The most-endorsed single post in the channel this year (28 hearts), and the clearest worked example of a group publishing its numbers. Raised by Yixiong Hao, Slack. A companion piece covers the prior year, LessWrong, Slack.
• The short timelines strategy for AI safety university groups (EA Forum). Josh Thorsteinson on what changes if you believe timelines are short. Raised by Josh Thorsteinson, Slack.
• Reflections on building an AI policy student group (Substack) and the accompanying workshop and presentation materials (Drive folder). The post is already cited on "Building a Strong Policy Group", the materials folder is not, and Liam deliberately made it public. Raised by Liam Robins, Slack.
• 10 lessons from running and mentoring AI safety groups (Substack). Manon Kempermann's writeup, explicitly aimed at organizers outside the US and UK. Raised by Manon Kempermann, Slack.