OpenAI publishes plan for ensuring AGI benefits everyone
Sam Altman and Jakub Pachocki set three goals for what they called OpenAI's third phase, including a personal AGI for every person and an automated AI researcher by March 2028.
- Ideas & essays
- Minor
OpenAI chief executive Sam Altman and chief scientist Jakub Pachocki co-authored an essay laying out what they called the company’s “third phase,” setting three explicit near-term goals: build an automated AI researcher capable of accelerating OpenAI’s own scientific work by March 2028, use AI to drive broad economic growth with widely distributed gains, and give every person on Earth access to a “personal AGI.”
A good AI future cannot be one where a small number of institutions control most of the capability.
The essay drew an explicit parallel to rural electrification in the 1920s, framing AGI as infrastructure whose benefit depends on wide distribution rather than concentration. It pointed to the OpenAI Foundation, the nonprofit that holds equity in OpenAI’s for-profit arm — then valued at roughly $130 billion — as a mechanism for that distribution, citing an initial $25 billion commitment across areas including health and disease research.
The document offered no binding commitments on governance structure, profit caps or timelines beyond the automated-researcher target, and read chiefly as a statement of intent rather than policy. It arrived amid continuing scrutiny of OpenAI’s own transition from nonprofit to a capped-profit, then more conventionally structured, company — a restructuring critics had argued cut against the “benefits humanity” language in OpenAI’s founding charter — making the essay as much a defence of that trajectory as a forward-looking plan.