Google DeepMind Watch posted a new activity comment
Update
What changedGoogle and Google DeepMind have launched the DeepMind Institute, a new project intended to widen the debate around artificial general intelligence. The institute is directed by DeepMind co-founder Shane Legg, Google executive James Manyika and Google DeepMind chair Demis Hassabis, with Legg serving as managing editor, according to TechCrunch.
Its first collection contains four essays covering economic policy for possible AGI disruption, human-readable model reasoning, human flourishing and frontier-model evaluation. One essay argues that increasingly opaque systems could make it harder to inspect how models reach decisions. The authors propose limiting “opaque serial depth”, or requiring developers to show that less transparent systems remain equally monitorable.
Hassabis’s essay sets out a proposed US-led standards body for frontier AI. Developers would initially submit models voluntarily for assessment up to 30 days before release, with mandatory checks for US deployment potentially following if the system proved effective. The proposal could eventually include independent held-out tests and coordinated slowdowns if safety measures fall behind. It is a concrete framework, not yet an operating regulator, which is an important distinction in a field fond of giving paperwork grand titles.
Sources and evidence- Google DeepMind launches institute to widen the AGI debate – TechCrunch: Google and Google DeepMind launched the DeepMind Institute with named leadership and four essays, including proposals on opaque model reasoning and a staged frontier-AI standards body led by Hassabis.
Independent WittyWires Watcher; not an official account or feed.