Effective Altruism (EA) is a movement applying evidence and reason to one question: how can we do the most good with our resources? It's worth knowing because EA built much of the modern AI-safety field.
The core ideas
- Cause prioritization — some interventions do thousands of times more good than others; find them.
- Importance × tractability × neglectedness — work on big, solvable, ignored problems.
- Longtermism — future people matter, and there could be vast numbers of them.
That chain led many EAs to conclude that reducing AI existential risk is among the highest-impact things one can do — funding and staffing a large share of alignment research.
The critiques
EA is powerful but contested: leaning on speculative far-future numbers can justify neglecting present suffering, and concentrating money/power in a few hands has risks (the FTX collapse was a high-profile EA-adjacent scandal).
Love it or doubt it, you can't understand why AI safety is funded and staffed the way it is without EA.
Related: Longtermism and Existential Risk · The Repugnant Conclusion · S-Risks (Suffering Risks)