每隔几代,就会有一项新技术改变一切。
想象一下 20 世纪 20 年代电力刚到达美国一个乡村小镇的情景。在电线架设之前,日常生活受制于各种物理条件的限制:挑水、手洗衣物、用冰块保存食物,太阳一落山,一天的许多活动便告结束。电力并没有在一夜之间改变每个家庭,它的许多惠及效应也是不均衡地抵达人们手中。但随着普及范围的扩大,普通人的生活发生了变化。夜晚的灯光延长了白天。电泵、电器和制冷设备减轻了一些最繁重的日常劳作。收音机把数百英里外的新闻、音乐和联系带进了家庭和社区空间。
电力的最初承诺是实用性的,但它更深远的影响来自更多人能够使用它之后所开启的全新可能。随着时间推移,大量新的可能性涌现出来,机器和计算机极大地加速了医学、工程学以及其他许多领域的进步。到 20 世纪末,人类平均寿命增加了 20 多年,经通胀调整后的收入中位数大约翻了三倍。这些成就的取得,在很大程度上得益于医疗、卫生和生活水平的进步,而其中许多进步正是由电气化的普及以及相关技术发展所促成或加速的。
这一切正在 AI 身上重演。AI 很快将能够完成非同寻常的事情。但关键不在于技术本身,而在于人们能用它做什么。它可以帮助人们处理医疗账单、学习新技能、创办小生意、照顾年迈的父母、理解法律或财务决策、把一个想法变成现实,或者做出科学发现。
虽然夜晚的灯光带来的惊奇可能很快就消退了,但人们决定用灯光去做什么却没有消退。而且由于技术长期以来一直是带来繁荣的可靠途径,我们认为 AI 应当向所有人开放,让人们可以在任何需要的地点、以任何需要的方式、按需尽情使用。
这样的未来不会自动到来。变革性技术既可能让权力集中,也可能让权力扩大范围。它们可以让少数人的生活更轻松,也可以为更多人拓展机会。我们的方法植根于这样一个信念:AI 应当服务于人——帮助人们追求自己的目标,增强他们的能力,并让这项技术的惠益尽可能广泛地分配。
我们的第一项承诺是构建服务于人类的 AI。这意味着我们希望广泛地赋能人们,而不是让权力集中在少数公司、政府或个人手中。我们相信,更安全的未来是权力广泛分布的未来,这样世界上更多的人才能参与建设一个有韧性的生态系统。*
我们对 AI 持乐观态度,因为我们相信它能够扩展人类的能力与繁荣。但我们也对风险有清醒的认识。强大的系统必须保持安全、与人类意图对齐,并处于人类的控制之下。我们在 OpenAI 的使命是确保 AGI 造福全人类。这意味着构建帮助人们做更多自己所选之事的系统,而不是构建取代人类判断何为重要的系统。
完全自动化一切并非我们想要的未来。那将令人空虚,也将十分危险。AI 应该帮助人们追求自己的目标,而不是脱离这些目标。随着 AI 系统能力越来越强,人的角色反而变得更加重要:设定方向、做出取舍、运用判断力,并将价值观、品味、关怀和责任感带入工作之中。
人们的一个关键长期角色,将是决定什么值得去做。
我们相信,在未来几年内,AI 开展 AI 研究将成为决定进步速度的关键因素。这一点很重要,因为对齐本身就是一项艰难的研究课题。要快速而深入地取得进展,我们的研究人员需要 AI 系统来帮助测试想法、发现错误、探索替代方案,并与我们并肩迭代。
但更快的技术进步使得人类判断和公共协调变得更加重要,而非更不重要。未来应当由人们、机构和社会共同塑造,而不仅仅由构建最强大系统的公司来决定。
随着前沿 AI 开发的持续推进,我们预计国家和全球层面的协调将变得更加重要。长期以来我们一直认为,最终应当建立一个国际组织,帮助协调各国的领先 AI 工作,以降低灾难性风险。合作与共同的安全标准是前进道路上的重要一环,尤其是因为商业竞争和国家竞争的驱动力难以摆脱。这样一个组织的目标之一,应当是让世界能够采取协调行动,包括在必要时放缓前沿开发,从而使社会的韧性、安全与对齐能够跟上步伐。
目前在 OpenAI,我们有三大主要目标
打造自动化的 AI 研究员——即一个能够加速并日益自动化研究过程本身的 AI 系统,同时保持可控、可问责,并与人类保持连接。我们的内部判断是,到 2028 年 3 月,我们可能有相当大比例的研究工作将由 AI 系统与我们自己的研究人员协同完成。为了在对齐方面取得足够进展,我们相信需要 AI 与我们共同迭代。这将帮助我们平稳过渡到后 AGI 世界,从而由我们共同决定通往未来的道路。
加速经济发展,通过加速科学进步、生产力和经济增长来实现,同时努力确保成果得到广泛共享。每个人都应有机会公平分享 AI 创造的繁荣。
让地球上的每个人都拥有一个专属的 AGI,使他们能够以自己选择的任何方式,从人类最具变革性的技术之一中受益。
为了实现这些目标,我们正在迈入 OpenAI 的第三阶段。
OpenAI 的第一阶段是开展面向 AGI 的研究。第二阶段始于我们的研究开始与现实世界产生关联、我们转型为一家产品公司之时:部署我们的系统,从人们的使用中学习,并在安全且符合我们使命的前提下,持续向 AGI 迈进。
如今我们正进入第三个阶段。经济格局正开始在 AI 周围重塑。现在的核心问题是,如何让先进的 AI 足够丰富、足够实惠、足够安全、足够有用、足够易用,让每个人和组织都能从中受益。前沿能力只是工作的一部分。更重大的任务是把这种能力转化为人们真正可以用来自身发展的工具。
最重要的是,我们相信权力的广泛分配将有助于通往更美好的未来。人类历史表明,权力集中会带来脆弱性,而广泛共享的权力则使社会更具韧性、适应力和自由。
这正是可及性如此重要的原因。这也是安全性、隐私、可负担性、开放生态和公众监督如此重要的原因。
一个美好的 AI 未来,绝不应是少数机构掌控大部分能力和大部分收益的未来。它应当是这样的未来:许许多多的人、企业、社区和国家都能够去构建、去受益、去掌握权力。我们相信,这场转型应当属于每一个人。
如果我们把这件事做对了,AI 就能成为提升生产力、创造力、科学进步以及普惠更多人的经济机遇的基石,我们也将实现我们的使命:确保 AGI 造福全人类。
*AI 韧性指的是社会可以建立起来的组织、系统和个体的集合,用于预判、承受、适应 AI 带来的冲击,并从中快速恢复。例如,汽车改变了社会,但它之所以能带来广泛的好处,是因为社会围绕它建立了一套体系:安全带、交通法规、驾驶执照、碰撞测试和道路基础设施。目标不是阻止人们开车,而是让一项强大的技术具备足够的韧性,从而得以广泛使用。
Every few generations, a new technology changes everything.
Imagine electricity reaching a rural American town in the 1920s. Before power lines arrived, daily life was shaped by physical limits: hauling water, washing clothes by hand, preserving food with ice, and ending much of the day when the sun went down. Electricity did not transform every household overnight, and many of its benefits reached people unevenly. But as access spread, ordinary life changed. Light at night extended the day. Electric pumps, appliances, and refrigeration reduced some of the hardest daily work. Radios brought news, music, and connection from hundreds of miles away into homes and community spaces.
The first promise of electricity was practical, but its deeper impact came from the new possibilities it opened as more people could use it. With time, a lot of new possibilities emerged, with machines and computers greatly accelerating progress in medicine, engineering, and many other fields. By the end of the 20th century the average lifespan had increased by over 20 years and the median inflation-adjusted income tripled or so. These gains were driven in no small part due to the advances in healthcare, sanitation, and living standards, many of which were enabled or accelerated by widespread electrification and related technological progress.
This is happening again with AI. AI will soon be capable of extraordinary things. But the point is not the technology by itself. The point is what people can do with it. It can help someone navigate a medical bill, learn a new skill, start a small business, care for an aging parent, understand a legal or financial decision, turn an idea into something real, or make a scientific discovery.
While the wonder of light at night probably wore off pretty fast, what people decided to do with it did not. And because technology has been a reliable way to deliver prosperity over time, we think AI should be available to everyone to use as much as they need, where and how they need it.
That future will not happen automatically. Transformative technologies can concentrate power, or they can broaden it. They can make life easier for a few, or they can expand opportunity for many. Our approach is rooted in the belief that AI should work for people: helping them pursue their own goals, increasing their capabilities, and distributing the benefits of this technology as widely as possible.
Our first commitment is to build AI in service of humanity. That means we want to empower people broadly, not see power concentrated among a few companies, governments, or individuals. We believe the safer future is one where power is broadly distributed, so more of the world can participate in building a resilience ecosystem.*
We are optimistic about AI because we believe it can expand human capability and prosperity. But we are also clear-eyed about the risks. Powerful systems must remain safe, aligned with human intent, and subject to human control. Our mission at OpenAI is to ensure that AGI benefits all of humanity. That means building systems that help people do more of what they choose, not systems that replace human judgment about what matters.
Entirely automating everything is not the future we want. It would be unfulfilling, and it would be dangerous. AI should help people pursue their goals, not become untethered from them. As AI systems become more capable, the human role becomes more important: setting direction, making tradeoffs, applying judgment, and bringing values, taste, care, and responsibility to the work.
A key long-term role for people will be deciding what is worth doing.
We believe that AI doing AI research will become the determining factor of the pace of progress within the next few years. That matters because alignment is itself a hard research problem. To make fast and deep progress, our researchers will need AI systems that can help test ideas, find mistakes, explore alternatives, and iterate alongside us.
But faster technical progress makes human judgment and public coordination more important, not less. The future should be shaped by people, institutions, and societies, not only by the companies building the most capable systems.
As frontier AI development continues, we expect national and global coordination to become more important. We have long believed there should ultimately be an international organization that helps coordinate leading AI efforts to reduce catastrophic risk. Cooperation and shared safety standards are an important part of the path forward, especially because the incentives around commercial and national competition are hard to escape. 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.
Currently at OpenAI we have three main goals
Build an automated AI researcher—an AI system that can accelerate and increasingly automate the research process itself, while remaining steerable, accountable, and connected to people. Our internal belief is 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. To make sufficient progress on alignment, we believe we will need AIs to iterate alongside us. This will help us navigate the transition to the post-AGI world so that we collectively decide the path toward the future.
Accelerate the economy, by accelerating scientific progress, productivity, and economic growth, while working to ensure the gains are widely shared. Everyone should have an opportunity for a meaningful share in the prosperity AI creates.
Give everyone on Earth a personal AGI, empowering them to benefit from one of humanity's most transformative technologies in whatever way they choose.
To be able to deliver on this, we are entering the third phase of OpenAI.
The first phase of OpenAI was about doing research toward AGI. The second phase began when our research became relevant to the real world and we became a product company: deploying our systems, learning from how people used them, and making continued progress toward AGI that is safe and aligned with our mission.
Now we are entering the third phase. The economy is beginning to reshape around AI. The central question now is how to make advanced AI abundant, affordable, safe, useful, and easy enough for every person and organization to benefit from it. Frontier capability is only part of the job. The bigger task is turning that capability into tools people can actually use to thrive.
Above all, we believe a broad distribution of power will help lead to a better future. Human history shows that concentrated power creates fragility, while widely shared power makes societies more resilient, adaptable, and free.
That is why access matters. It is also why safety, privacy, affordability, open ecosystems, and public oversight matter.
A good AI future cannot be one where a small number of institutions control most of the capability and most of the upside. It should be a future where many people, companies, communities, and countries can build, benefit, and hold power. We believe this transformation should belong to everyone.
If we get this right, AI can become a foundation for greater productivity, creativity, scientific progress, and economic opportunity for the many, and we will achieve our mission: to ensure AGI benefits all of humanity.
*AI resilience refers to the collective organizations, systems, and individuals that society could put in place to anticipate, withstand, adapt to, and rapidly recover from AI-driven disruptions. For instance, the automobile transformed society, but they only became broadly beneficial because societies built systems around them: seatbelts, traffic laws, drivers’ licenses, crash testing, and road infrastructure. The goal wasn’t to stop people from driving—it was to make a powerful technology resilient enough for widespread use.