# 演进 OpenAI 的结构(Evolving OpenAI's structure) OpenAI 董事会已更新了关于演进 OpenAI 结构的计划。 OpenAI 成立时是一家非营利组织,如今由该非营利组织监督和控制。展望未来,它将继续由该非营利组织监督和控制。 我们的营利性有限责任公司(LLC)自 2019 年以来一直隶属于该非营利组织,将转型为公益公司(Public Benefit Corporation, PBC)——这是一种以使命为导向的公司结构,必须同时考虑股东利益与使命。 该非营利组织将控制 PBC,同时成为其大股东,从而使该非营利组织拥有更好的资源来支持诸多公益。 我们的使命保持不变,PBC 也将拥有同样的使命。 在听取公民领袖的意见,并与特拉华州总检察长办公室和加利福尼亚州总检察长办公室进行了建设性对话之后,我们做出了由非营利组织保留对 OpenAI 控制权的决定。我们感谢这两个办公室,并期待继续这些重要对话,以确保 OpenAI 能够继续有效追求其使命——确保 AGI 造福全人类。Sam 在下面写了一封信,致我们的员工和利益相关方,说明我们为何对这一新方向感到如此兴奋。——Bret Taylor * * * The OpenAI Board has an updated plan for evolving OpenAI’s structure. ¶
## Sam 致员工的信 OpenAI 不是一家普通公司,也永远不会是。 我们的使命是确保人工通用智能(AGI)造福全人类。 当我们创办 OpenAI 时,对于如何实现使命并没有一个详细的概念。我们起初围坐在一张餐桌旁面面相觑,思索应该做些什么研究。那时,我们并未考虑产品、商业模式。我们无法想象 AI 被用于医疗建议、学习、生产力等诸多领域所带来的直接益处,也无法想象训练模型和服务用户所需的数千亿美元算力。 我们并不真正知道 AGI 将如何被构建,或如何被使用。很多人可以想象一个能告诉科学家和总统该做什么的神谕,尽管它可能极其危险,但也许少数这样的人是可以被信任的。 早期在 OpenAI 周围的很多人认为,AI 应该只掌握在少数能够「驾驭它」的可信之人手中。 如今,我们看到了一条道路,让 AGI 作为人类历史上最强大的工具,直接赋能每一个人。如果我们能做到这一点,我们相信人们将为彼此创造不可思议的东西,并持续推动社会和生活质量的进步。当然,它不会全部被用于善途,但我们相信人类,并认为善将比恶多出数个数量级。 我们致力于这条民主化 AI 的道路。我们希望把不可思议的工具交到每个人手中。我们对人们用我们的工具所创造的东西,以及他们使用这些工具的强烈意愿,感到惊讶与欣喜。我们希望开源非常强大的模型。我们希望在宽泛的边界内,给予用户在如何使用我们工具方面极大的自由,即便我们并不总是共享同一套道德框架,并让用户就 ChatGPT 的行为做出决定。 我们相信这是最好的前进道路——AGI 应当使全人类能够彼此受益。我们意识到有些人的看法非常不同。 我们希望为世界构建一个大脑,并让人们能够极其便利地将它用于任何他们想做的事情(受少数限制约束;例如,自由不应侵犯他人的自由)。 人们正在使用 ChatGPT 提高其作为科学家、程序员等的生产力。人们正在使用 ChatGPT 解决他们所面临的严肃医疗挑战,并以前所未有的程度学习。人们正在使用 ChatGPT 获取关于如何应对困难处境的建议。我们非常自豪能够提供这样一项为如此多的人做如此多事情的服务;这是我们能想象到的对使命最直接的实现之一。 但他们想要更多地使用它;我们目前无法提供世界所想要的近乎足量的 AI,我们不得不对系统设置使用限制,并让其缓慢运行。随着系统变得更加强大,人们会想要更多地使用它,用于更美好的事情。 大约十年前我们启动研究实验室时,完全没想到世界会是这样的状态。但现在我们看到了这幅图景,我们感到无比振奋。 是时候演进我们的结构了。我们想要完成三件事: 我们希望以一种能够让我们向全人类广泛提供服务的方式来运营并获取资源,而这目前需要数千亿美元,最终可能需要数万亿美元。我们相信这是我们实现使命、并让人们用这些新工具为彼此创造巨大利益的最佳方式。 我们希望我们的非营利组织成为历史上规模最大、最有效的非营利组织,专注于利用 AI 为人们创造最高杠杆的成果。 我们希望交付有益的 AGI。这包括为安全与对齐的形态贡献力量;我们为已发布系统的记录、所做的对齐研究、红队测试(red teaming)等流程,以及通过模型规范(model spec)等创新实现的模型行为透明度感到自豪。随着 AI 加速发展,我们对安全的承诺愈发坚定。我们希望确保民主的 AI 战胜威权的 AI。 在听取公民领袖的意见并与加利福尼亚州和特拉华州总检察长办公室讨论后,我们做出了由非营利组织保持控制的决定。我们期待在与他们、微软以及我们新任命的非营利专员的持续对话中,推进这一计划的细节。 OpenAI 成立时是一家非营利组织,如今是一家监督和控制营利实体的非营利组织,未来仍将是一家监督和控制营利实体的非营利组织。这一点不会改变。 隶属于非营利组织的营利性 LLC 将转型为一家拥有相同使命的公益公司(PBC)。PBC 已成为其他 AGI 实验室(如 Anthropic 和 X.ai)以及诸多使命驱动型公司(如 Patagonia)的标准营利结构。我们认为这对我们也是合理的。 我们将摆脱目前复杂的利润上限结构——当看起来可能只有一项占主导地位的 AGI 努力时,这种结构是合理的,但在一个拥有许多优秀 AGI 公司的世界里则不然——转而采用一种每个人都能持有股份的普通资本结构。这不是出售,而是向更简单结构的转变。 非营利组织将继续控制 PBC,并将成为 PBC 的大股东,其持股规模由独立财务顾问支持,使非营利组织拥有资源来支持相关项目,从而让 AI 能够造福多个不同社区,与使命保持一致。随着 PBC 的成长,非营利组织的资源也将增长,从而能够做得更多。我们很高兴很快能收到非营利委员会关于如何确保 AI 造福每个人——而非少数人——的建议。他们的想法将聚焦于我们的非营利工作如何支持一个更加民主的 AI 未来,并在健康、教育、公共服务和科学发现等领域产生真实影响。 我们相信,这使我们能够继续快速、安全地取得进展,并把强大的 AI 交到每个人手中。创造 AGI 是我们在人类进步之路上的一块砖;我们迫不及待想看到你们接下来添上的砖。 Sam Altman 2025 年 5 月 【译注:原文此后为网页导航壳层内容(2025、Authors、Research Index、GPT 型号列表、ChatGPT 各版本、API 登录等站点链接),非正文,依项目惯例不译。】 OpenAI was founded as a nonprofit, and is today overseen and controlled by that nonprofit. Going forward, it will continue to be overseen and controlled by that nonprofit. Our for-profit LLC, which has been under the nonprofit since 2019, will transition to a Public Benefit Corporation (PBC)–a purpose-driven company structure that has to consider the interests of both shareholders and the mission. The nonprofit will control and also be a large shareholder of the PBC, giving the nonprofit better resources to support many benefits. Our mission remains the same, and the PBC will have the same mission. We made the decision for the nonprofit to retain control of OpenAI after hearing from civic leaders and engaging in constructive dialogue with the offices of the Attorney General of Delaware and the Attorney General of California. We thank both offices and we look forward to continuing these important conversations to make sure OpenAI can continue to effectively pursue its mission of ensuring AGI benefits all of humanity. Sam wrote the letter below to our employees and stakeholders about why we are so excited for this new direction.—Bret Taylor * * * Sam’s Letter to Employees. OpenAI is not a normal company and never will be. Our mission is to ensure that artificial general intelligence (AGI) benefits all of humanity. When we started OpenAI, we did not have a detailed sense for how we were going to accomplish our mission. We started out staring at each other around a kitchen table, wondering what research we should do. Back then, we did not contemplate products, a business model. We could not contemplate the direct benefits of AI being used for medical advice, learning, productivity, and much more, or the needs for hundreds of billions of dollars of compute to train models and serve users. We did not really know how AGI was going to get built, or used. A lot of people could imagine an oracle that could tell scientists and presidents what to do, and although it could be incredibly dangerous, maybe those few people could be trusted with it. A lot of people around OpenAI in the early days thought AI should only be in the hands of a few trusted people who could “handle it”. We now see a way for AGI to directly empower everyone as the most capable tool in human history. If we can do this, we believe people will build incredible things for each other and continue to drive society and quality of life forward. It will of course not be all used for good, but we trust humanity and think the good will outweigh the bad by orders of magnitude. We are committed to this path of democratic AI. We want to put incredible tools in the hands of everyone. We are amazed and delighted by what they are creating with our tools, and how much they want to use them. We want to open source very capable models. We want to give our users a great deal of freedom in how we let them use our tools within broad boundaries, even if we don’t always share the same moral framework, and to let our users make decisions about the behavior of ChatGPT. We believe this is the best path forward—AGI should enable all of humanity to benefit each other. We realize some people have very different opinions. We want to build a brain for the world and make it super easy for people to use for whatever they want (subject to few restrictions; freedom shouldn’t impinge on other people’s freedom, for example). People are using ChatGPT to increase their productivity as scientists, coders, and much more(opens in a new window). People are using ChatGPT to solve serious healthcare challenges they are facing and learn more than ever before. People are using ChatGPT to get advice about how to handle difficult situations. We are very proud to offer a service that is doing so much for so many people; it is the one of most direct fulfillments of our mission we can imagine. But they want to use it much more; we currently cannot supply nearly as much AI as the world wants and we have to put usage limits on our systems and run them slowly. As the systems become more capable, they will want to use it even more, for even more wonderful things. We had no idea this was going to be the state of the world when we launched our research lab almost a decade ago. But now that we see this picture, we are thrilled. It is time for us to evolve our structure. There are three things we want to accomplish: We want to be able to operate and get resources in such a way that we can make our services broadly available to all of humanity, which currently requires hundreds of billions of dollars and may eventually require trillions of dollars. We believe this is the best way for us to fulfill our mission and to get people to create massive benefits for each other with these new tools. We want our nonprofit to be the largest and most effective nonprofit in history that will be focused on using AI to enable the highest-leverage outcomes for people. We want to deliver beneficial AGI. This includes contributing to the shape of safety and alignment; we are proud of our track record with the systems we have launched, the alignment research we have done, processes like red teaming, and transparency into model behavior with innovations like the model spec(opens in a new window). As AI accelerates, our commitment to safety grows stronger. We want to make sure democratic AI wins over authoritarian AI. We made the decision for the nonprofit to stay in control after hearing from civic leaders and having discussions with the offices of the Attorneys General of California and Delaware. We look forward to advancing the details of this plan in continued conversation with them, Microsoft, and our newly appointed nonprofit commissioners. OpenAI was founded as a nonprofit, is today a nonprofit that oversees and controls the for-profit, and going forward will remain a nonprofit that oversees and controls the for-profit. That will not change. The for-profit LLC under the nonprofit will transition to a Public Benefit Corporation (PBC) with the same mission. PBCs have become the standard for-profit structure for other AGI labs like Anthropic and X.ai, as well as many purpose driven companies like Patagonia. We think it makes sense for us, too. Instead of our current complex capped-profit structure—which made sense when it looked like there might be one dominant AGI effort but doesn’t in a world of many great AGI companies—we are moving to a normal capital structure where everyone has stock. This is not a sale, but a change of structure to something simpler. The nonprofit will continue to control the PBC, and will become a big shareholder in the PBC, in an amount supported by independent financial advisors, giving the nonprofit resources to support programs so AI can benefit many different communities, consistent with the mission. And as the PBC grows, the nonprofit’s resources will grow, so it can do even more. We’re excited to soon get recommendations from our nonprofit commission on how we can help make sure AI benefits everyone—not just a few. Their ideas will focus on how our nonprofit work can support a more democratic AI future, and have real impact in areas like health, education, public services, and scientific discovery. We believe this sets us up to continue to make rapid, safe progress and to put great AI in the hands of everyone. Creating AGI is our brick in the path of human progress; we can’t wait to see what bricks you will add next. ¶