我们正站在一场深刻变革的门槛上。今天的 AI 系统能力惊人，创新的飞快节奏表明，我们正在接近通用人工智能（AGI），也就是一个展现出人脑全部认知能力的系统。虽然 AI 有时仍会在基础任务上失败，也还缺乏达到完整 AGI 所需的稳定性和创造力，但我们预计这些差距很快会被弥合。我们一直相信，AGI 可能成为加速科学突破的终极工具。通过开辟理解和治愈疾病的新路径、开发清洁能源、提升经济繁荣，它可能开启一个科学发现与进步的新黄金时代，远远超出没有它时我们所能达到的程度。
但这样一项变革性技术的到来，也伴随着挑战和风险。我们已经看到，能力越来越强的 AI 对网络安全和生物风险带来的影响；未来能够自我改进的系统，还有失控的可能。现在是一个关键时刻，要确保我们安全地构建 AGI，并让它给社会带来的好处远远超过任何风险。
DeepMind's own institute: debater, referee and venue in one, landing in a crowd already allergic to labs building their own governance.
How to read this
Two layers of evidence of different strength: the launch text is a first-hand original on X and can be quoted; the X reactions and engagement figures come from search and are second-hand, still to be checked. The launch itself is heavily self-interested: DeepMind pulls the say over AGI governance onto a platform it hosts, and enlists the arts, the humanities and governments as endorsement.
Shane Legg says he has worked toward AGI for 25 years; AGI is on the horizon, so DeepMind has created the institute to deepen understanding of what it will bring.
My journey to develop AGI spans 25 yrs, including 10+ yrs thinking about technical & societal perspectives at Google DeepMind.
AGI is on the horizon - we need deeper understanding of its implications. To help, we've created the DeepMind Institute.
02
The AGI claim: warmer in the launch text than from Legg himself
AGI is defined as a system with all the cognitive capabilities of the human brain, and today's gaps are expected to close “soon”; cybersecurity, biorisk and loss of control in future self-improving systems are acknowledged.
We are on the cusp of a profound transformation. Today’s AI systems have impressive capabilities and the rapid pace of innovation suggests we’re now approaching artificial general intelligence (AGI), a system that exhibits all the cognitive capabilities of the human brain. While AI can still sometimes fail at basic tasks and lacks the consistency and creativity to meet the bar of full AGI, we expect those gaps to be closed soon. We’ve always believed AGI could prove to be the ultimate tool for accelerating scientific breakthroughs. By opening up new paths to understanding and curing disease, developing clean energy and increasing economic prosperity, it could unlock a new golden age of scientific discovery and progress far beyond what we could achieve without it.
Yet there are also challenges and risks that come with the advent of such a transformative technology. We’re already seeing the implications of increasingly capable AI for cybersecurity and biorisks, and there is the potential for loss of control in future self-improving systems. This is a critical moment to ensure we build AGI safely and its benefits to society far outweigh any risks.
03
Three governance questions, and “responsibility belongs to society”
DMI is an interdisciplinary publishing platform, not a product or a regulator. It asks what we'll value, how to build and govern AGI safely, and which institutions need rethinking, and says shaping AGI belongs to society as a whole.
We’re launching the DeepMind Institute (DMI) to spur the interdisciplinary research, collaboration and debate required to answer the AGI era's most critical technical and societal questions. What will we value, and how will AGI impact what it means to be human? How do we safely build and govern AGI systems and the communities of agents they will form? Which institutions and policies will society need to adapt to AGI or reimagine altogether?
DMI is a platform for researchers and thinkers from across Google DeepMind, Google, and the wider global research community to work on and publish creative, deeply informed ideas about a world with AGI. They will not always agree, and they will likely change their minds, as more data and information comes to light at the fast-moving frontier. We created this institute precisely because broad-based intellectual discussion and debate are required to arrive at a consensus about how to address the challenges and opportunities we face as a society.
Nobody has all the answers about how AGI will be developed and deployed responsibly in the world, and it shouldn’t be technologists alone who provide them. Technological progress and expertise establish only the possibility of a new era; the responsibility of shaping its reality collectively belongs to society as a whole, including the arts and humanities, and governments. DMI brings diverse viewpoints together to identify the critical challenges we need to tackle, debate the potential solutions, and help ensure AGI improves the lives of everyone. We look forward to these discussions, which are essential as society thinks through how to safely steward AGI into the world.
DeepMind Institute Directors: Demis Hassabis, James Manyika, Shane Legg
Where Indigo landsFurther
Indigo's conclusion
A soft positioning move by Google on the AGI narrative and the say over governance: a controlled discussion hosted by the builder. Debater, referee and venue in one; it isn't independent and can't replace outside oversight.
What to remember
Separate criticism of DMI from the crowd it landed in: the day's fire was aimed mainly at METR and the slow-down push; DMI was just the newest example.
On X, the official accounts' own reposts dominated; the sharpest criticism sat in low-engagement replies and hasn't become a shareable counter-narrative.
Outside researchers' default stance: read it like a journal first, and see whether it can publish anything inconvenient for a launch window.
DMI is narrative capital, set against Anthropic's institute; one handshake emoji from Jack Clark and the two labs recognized each other's legitimacy.
Back on the long-running theses
confirms
The politics of open-weight safety Every side's safety framework happens to reinforce its own position, and DMI is another case; two labs' institutes recognizing each other is the club inside the gate protecting itself.
conflicts
Dario Amodei, We must slow the frontier DMI is self-regulation with the company choosing what to disclose; Dario wants third parties with employee-level access. Two different routes.
A soft asset for Google, not a fundamentals event: no product, no revenue meaning, no near-term impact. What matters is narrative capital, who defines AGI and who hosts the governance debate, and the two big labs vouching for each other's legitimacy.
What would change my mind
DMI publishing a paper that is inconvenient for a Google launch window, or that raises a real, independent objection on frontier-model safety.
Finished. Indigo's take on this piece is in two places: