Google DeepMind launches institute to study life after AGI
Google DeepMind has launched a new institute to study how AGI could reshape society, science and the economy, bringing internal and external researchers into the debate.
AGI is no longer only a question for AI researchers. Google DeepMind has launched the DeepMind Institute, a new platform focused on exploring how artificial general intelligence could affect economies, science and society.
The institute brings together researchers from Google DeepMind and external organisations to examine some of the questions that could become more important as AI systems become more capable.
Looking beyond AI capabilities
Artificial general intelligence, or AGI, generally refers to AI systems capable of performing a broad range of cognitive tasks at around human level or beyond. The DeepMind Institute is not a new model or AI product. Instead, it is a platform for publishing research and essays on what AGI could mean for society.
Google DeepMind says the initiative is intended to bring different perspectives together and make informed discussion around AGI more accessible. Its initial work covers areas including economic policy, transparency in AI reasoning, global access to advanced AI and human well-being.
That makes the focus broader than model performance. The institute is also looking at questions around who benefits from advanced AI, how its effects could be managed and what society may need to prepare for.
Shane Legg leads the institute
The institute is directed by James Manyika, Shane Legg, and Demis Hassabis, with Legg serving as managing editor. Legg is co-founder and Chief AGI Scientist at Google DeepMind, while Manyika is a senior Google executive and Hassabis is the CEO and co-founder of DeepMind.
Legg has long worked on questions around artificial intelligence and AGI. Google DeepMind's own research history also reflects this focus, with the company describing AGI as part of its longer-term vision for AI research.
The institute's work is presented as a forum for discussion rather than an official statement of Google's position. Its website explicitly says that individual pieces reflect their authors' ideas and research.
From research to public debate
One of the institute's early essays examines economic policy for a future in which AGI could significantly affect jobs and productivity. Another looks at transparency in AI reasoning, while other work explores access to advanced AI and broader questions about human flourishing.
The launch comes as questions around AI governance are becoming increasingly international. Issues such as access, economic disruption, safety and the development of common standards extend beyond the companies building the technology.
For countries such as India, these discussions could eventually intersect with questions around skills, jobs, access to AI and how the benefits of increasingly capable systems are distributed.
The DeepMind Institute does not mean AGI has arrived. Its launch instead reflects a growing effort to examine the social and economic questions around advanced AI alongside the technical work happening inside AI labs.