The Pulse
DeepMind Institute Introduces Forum for AGI Questions
The DeepMind Institute says it will bring researchers and thinkers together to examine the technical, social and governance questions surrounding artificial general intelligence.
AI.info Team ·
“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.”
DeepMind Institute launch essay; Shane Legg (director and managing editor), James Manyika (director) and Demis Hassabis (director) are listed on the page
A Forum for AGI Questions
The DeepMind Institute says it will advance research, collaboration and debate about artificial general intelligence and its implications for society.
The institute is directed by Shane Legg, James Manyika and Demis Hassabis. Legg also serves as managing editor. The page presents the institute as a platform for researchers and thinkers from Google DeepMind, Google and the wider global research community to develop and publish ideas about a world with AGI.
Its stated mission combines technical questions about the safe development of AGI with wider questions about how such systems could be used and governed. The institute says its work will examine the benefits AGI might bring, as well as the risks and social changes that could accompany increasingly capable systems.
From Model Safety to Human Values
The launch essay describes AGI as a system that exhibits all the cognitive capabilities of the human brain. It says current AI systems can still fail at basic tasks and lack the consistency and creativity associated with full AGI, while arguing that those gaps may close soon.
The institute's agenda extends beyond model performance. Its launch material asks what people will value as AGI develops, how such systems and communities of agents can be built and governed safely, and which institutions and policies society may need to adapt or rethink.
The essay also argues that technologists should not be the only people responsible for deciding how AGI is developed and deployed. It calls for broad participation, including contributions from the arts and humanities, governments and other parts of civil society.
The institute says that broad intellectual discussion is needed to address the challenges and opportunities associated with AGI. Its contributors may disagree with one another and may change their positions as new information becomes available.
A Publication and Discussion Platform
The DeepMind Institute does not announce a new model or technical system in its launch essay. Instead, it establishes a venue for publishing and discussing ideas about the future of AGI.
The institute's disclaimer says its pieces are intended as conversation starters that reflect the ideas and research of their authors. They should not be read as Google's official view. That position allows the platform to host differing arguments while separating those discussions from formal corporate policy.
The launch essay says nobody has all the answers about how AGI should be developed and deployed responsibly. It presents the responsibility for shaping that future as a collective one, involving society as a whole rather than technologists alone.
With Legg, Manyika and Hassabis serving as directors, the DeepMind Institute begins as a publication and discussion platform focused on the questions that may arise as AI systems approach broader human capabilities. Its work is intended to bring different fields and viewpoints into that conversation before agreement exists on the answers.