OpenAI Charter

The OpenAI Charter (April 2018, updated) is the foundational public commitment document that governs OpenAI's decision-making — and serves as the primary accountability mechanism that external critics and internal employees use to challenge the company's choices. Four core commitments structure the charter: broadly beneficial AGI (benefiting all of humanity, not OpenAI shareholders or any subset of people); long-term safety (halting or adjusting deployment if safety risks emerge); technical leadership (being at the frontier because a safety-focused organization needs to understand the technology); and cooperative orientation (working with other institutions rather than racing unilaterally). The charter is significant because it explicitly states OpenAI will abandon its own mission if another project appears to be building safe AGI and OpenAI can contribute more by supporting them than competing. Critics have noted the gap between the charter's stated commitments and OpenAI's commercial behavior: the transition to a capped-profit company, the Microsoft partnership, and accelerating deployment schedules are all difficult to reconcile with "long-term safety" as the primary priority. Altman's defense has been consistent: safety and capabilities research are complementary, not competing, because you can only build safe systems if you understand the most capable systems. The charter is the lens through which every major OpenAI decision — o1, GPT-4, Sora, the governance crisis — should be read.