cd /news/ai-safety/ais-collab-is-piloting-an-agi-prepar… · home topics ai-safety article
[ARTICLE · art-40701] src=forum.effectivealtruism.org ↗ pub= topic=ai-safety verified=true sentiment=· neutral

AIS Collab is Piloting an AGI Preparedness Course: The Better Futures Fellowship

AIS Collab is piloting a nine-week Better Futures Fellowship program aimed at individuals deciding whether to focus on research and fieldbuilding to improve the quality of future civilizations, conditional on humanity surviving. The program combines five weeks of facilitated discussion based on William MacAskill's research at Forethought Institute with a four-week project phase where each fellow produces a public output. The fellowship targets people familiar with effective altruism and AI safety who are open to pivoting on neglectedness grounds.

read3 min views1 publishedJun 26, 2026

TL;DR: We're opening an expression of interest for the Better Futures Fellowship pilot. This is a 9-week, part-time program for people deciding whether better-futures work— * * research and fieldbuilding aimed at improving how good the future is, conditional on humanity surviving — should become their focus. Five weeks of facilitated discussion grounded in research by William MacAskill and his colleagues at Forethought Institute, followed by a four-week project phase where every fellow ships a public output. Cohorts of 4–8. Submit your expression of interest here.

I used AI to assist in writing this post, and it’s likely that >30% is AI-generated text.

The Better Futures Fellowship pilot is a nine-week, part-time program for people deciding whether better-futures work — research and fieldbuilding aimed at improving how good the future is, conditional on humanity surviving — should become their focus. It combines a five-week facilitated discussion course anchored in research by William MacAskill and his colleagues at Forethought Institute, with a four-week guided project phase in which every fellow produces a concrete research or fieldbuilding output.

Most longtermist effort goes into making sure there is a future at all. The Better Futures series argues that this is only half the problem: if futures where we survive realise just a small fraction of what is possible, then improving the quality of surviving futures may matter as much as raising the odds of survival (by some estimates far more) and it is dramatically more neglected. The field working on this is currently a handful of researchers. The fellowship exists to grow it.

The program builds on the infrastructure of AIS Collab. **Primary audience: **people familiar with effective-altruism principles who have engaged seriously with cause prioritisation, currently in or considering AI safety and who are open to pivoting on neglectedness grounds.

**Secondary audience: **people already engaged with AI safety (e.g. BlueDot graduates) without an EA background, comfortable with existential risk and transformative AI concepts. We point to EA foundations (impartiality, longtermism, value lock-in) in optional readings where needed.

**Not for: **people without prior EA or AI-safety familiarity. The course assumes that context from day one.

**Selection emphasis: **motivated and agentic 'founder-profile' people, strong generalists, and policy entrepreneurs. The field is emerging, so fellows must be comfortable working without established career tracks.

Wk Theme Core reading Exercise Session centerpiece
1 Why flourishing, not just survival ~53 min Your baseline picture + strongest objection Two Futures; the 36x stress-test
2 The intelligence explosion and grand challenges ~63 min Pick your grand challenge; apply the punt test Concept teach-backs; punt-or-can't-punt sorting
3 What would a good future actually be? ~49 min Mistopia sketch Mistopia clinic; designing for moral uncertainty
4 Power concentration: who ends up controlling AGI? ~55 min Threat-model a concentration pathway Kill-chain red-teaming; coup-proofing an AGI project
5 AI character, epistemics, and choosing your lever ~57 min Two project pre-proposals Project speed-dating; direction commitment
6–9 Project phase (see Project Phase Guide) Milestones M1–M4 Proposal workshop → midpoint clinic → peer review → showcase

This is an expression of interest, not a full application. IThis course is based on research by William MacAskill and his colleagues at Forethought Institute.t should take about 3 minutes. We'll follow up with you once applications open.

Submit your expression of interest

This curriculum draws on Forethought's published research program on better futures, and was developed with feedback from Forethought's team. If you're interested in following the underlying research rather than the fellowship itself, see Forethought's Better Futures series.

── more in #ai-safety 4 stories · sorted by recency
── more on @ais collab 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/ais-collab-is-piloti…] indexed:0 read:3min 2026-06-26 ·