Primary text: openai.com Authors: Sam Altman and Jakub Pachocki, OpenAI. Published June 8, 2026.
"Built to benefit everyone: our plan" is a company post published June 8, 2026 by Sam Altman and OpenAI chief scientist Jakub Pachocki, setting out OpenAI's strategy for what it calls its third phase. The post argues that transformative technologies can concentrate power or broaden it, that "the safer future is one where power is broadly distributed," and that OpenAI's mission — ensuring AGI benefits all of humanity — means building systems that augment human judgment rather than replace it. It opens with an analogy to rural electrification in the 1920s, arguing the deeper value of a general-purpose technology comes from what people do with it as access spreads.
Three goals
The post states three current OpenAI goals:
- Build an automated AI researcher — a system that can accelerate and increasingly automate the research process while remaining steerable and accountable. The authors state an internal belief that "by March of 2028 we may have a significant fraction of our research being done by AI systems in tandem with our own researchers," and argue alignment is itself a hard research problem that will require AI systems to iterate alongside human researchers.
- Accelerate the economy — through faster scientific progress, productivity, and growth, "while working to ensure the gains are widely shared."
- Give everyone on Earth a personal AGI — empowering people to benefit from the technology in whatever way they choose.
The "third phase"
The post frames OpenAI's history in three phases: a first phase of AGI research; a second phase, beginning when that research became commercially relevant, as a product company deploying systems and learning from use; and a third phase, now beginning, defined by the question of how to make advanced AI "abundant, affordable, safe, useful, and easy enough for every person and organization to benefit from it." The authors argue frontier capability is "only part of the job" and the larger task is turning capability into usable tools. A recurring thesis is that "a broad distribution of power will help lead to a better future," that concentrated power "creates fragility," and that a good AI future "cannot be one where a small number of institutions control most of the capability and most of the upside."
Safety, human control, and international coordination
The authors state that powerful systems "must remain safe, aligned with human intent, and subject to human control," and that "entirely automating everything is not the future we want," arguing the human role becomes more important as systems become more capable — setting direction, making tradeoffs, and applying judgment. They write that OpenAI has "long believed there should ultimately be an international organization that helps coordinate leading AI efforts to reduce catastrophic risk," and that one goal of such an organization "should be to make it possible for the world to take coordinated action, including slowing frontier development when needed, so societal resilience, safety, and alignment can keep pace." The post introduces "AI resilience" as a framing — defined in a footnote as the organizations, systems, and individuals society could put in place to anticipate, withstand, adapt to, and recover from AI-driven disruptions — with an analogy to the seatbelts, traffic laws, licensing, crash testing, and road infrastructure built around the automobile.
Reception and context
The post was published days before OpenAI's June 8, 2026 confidential S-1 filing with the SEC and amid reporting that OpenAI expected to go public "within the next year" (Source: openai.com; theinformation.com). Coverage characterized it as OpenAI's "third phase" vision and quoted the "slowing frontier development when needed" passage (Source: insideaipolicy.com). Its emphasis on broad distribution of power and an international coordinating organization contrasts with the same-week Amodei essay (June 10, 2026), which proposes binding FAA-style national testing of frontier models with government authority to block or reverse deployment. As a company position statement, it is treated as an argument rather than as factual evidence.
Provenance
Company post on openai.com, June 8, 2026, co-authored by OpenAI's CEO and chief scientist. The raw file was pulled and authenticity-verified on the canonical host during the June 11, 2026 gap scan, corroborated by Inside AI Policy, TNW, and AOL/Business Insider coverage and the New Developments Log entry of June 10, 2026.
Relationships
- related: OpenAI; Sam Altman; Artificial General Intelligence (AGI); AGI Timelines; Recursive Self-Improvement (RSI); Frontier AI Governance.
- contradicts: (partial) Policy on the AI Exponential (Dario Amodei, June 2026) — emphasizes broad distribution and an international coordinating organization over binding national testing and deployment-blocking authority.