Google DeepMind Launches Institute to Broaden AGI Debate
Artificial general intelligence, or AGI, could reshape labor markets, corporate competition and public governance, but researchers still disagree on how it should be defined and how quickly it may arrive. The DeepMind Institute, launched by researchers from Google and Google DeepMind, is intended as an interdisciplinary forum for examining those economic and social consequences while broadening debate over model safety, transparency and oversight beyond the companies building frontier systems.
Google DeepMind announced the institute on Sept. 16, 2026, naming Shane Legg, James Manyika and Demis Hassabis as directors, with Legg also serving as managing editor; no budget was disclosed. Its first four essays address economic policy for AGI disruption, human-readable model reasoning, human flourishing and evaluation of frontier AI. Legg reiterated his estimate of roughly a 50% chance that “minimal AGI” will emerge by 2028, arguing that safeguards need to advance alongside technical capabilities.
All Coverage
2 original reportsThe Backstory
The history behind this eventGoogle DeepMind Report Maps Pathways and Challenges From AGI to ASI
Google DeepMind has released a 57-page report arguing that artificial general intelligence, or AGI, is no longer sufficient to describe the endpoint of frontier AI development. It instead uses artificial superintelligence, or ASI, as its central framework. The report defines the ASI threshold as capabilities exceeding what tens of thousands of top experts could achieve by working together for a decade, with implications for technological competition, governance and the future of human decision-making authority.
As of July 20, 2026, coverage has focused on DeepMind’s four potential pathways to ASI, as well as six inherent advantages AI could possess and six barriers to its long-term development. Media reports have translated the capability threshold into the collective output of about 100 million average people, underscoring that ASI would represent more than an improvement in model performance and could shape technological and societal trajectories for decades.
Subscribe to Mark Radar Weekly
Every Friday, the week's strongest signals in your inbox. Unsubscribe anytime.
If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →