Google Deepmind launched the Deepmind Institute (DMI) on September 16, 2026, to address the safety and governance of artificial general intelligence (AGI). The institute aims to bring together researchers from Google Deepmind, Google, and the global scientific community to explore AGI's risks and implications.
The institute emphasizes collaboration across disciplines, including technology, arts, humanities, and policy, according to directors Shane Legg, James Manyika, and Nobel laureate Demis Hassabis. They argue that AGI's challenges require input beyond technologists to ensure comprehensive understanding and responsible development.
There is no agreed-upon definition of AGI, but Deepmind describes it as a system with all the cognitive abilities of the human brain. Despite current AI systems failing at simple tasks and lacking creativity, Hassabis believes AGI could emerge within a few years, with a precursor possibly arriving by 2028.
"AGI shouldn't be confused with 'ASI', an artificial superintelligence that would far surpass humans in every respect," said Demis Hassabis, a Nobel laureate and director of the institute. "AGI is likely a necessary step toward ASI, but the focus remains on understanding and managing its risks.
The announcement follows growing interest in AGI safety, with OpenAI focusing on economic value and CEO Sam Altman predicting AGI could arrive by year's end. The institute's launch reflects a broader industry effort to address AGI's potential impacts before its arrival.
Google Deepmind did not specify a timeline for AGI's development and raised questions about its risks and governance. The institute will continue to explore these issues through interdisciplinary research and debate.
Source: thedecoder
