Google DeepMind launches institute to widen the AGI debate - TechCrunch

Google and Google DeepMind researchers launched the DeepMind Institute on Wednesday, according to TechCrunch, with the stated aim of advancing the conversation around artificial general intelligence. The institute lists three directors: DeepMind co-founder Shane Legg, Google executive James Manyika and Google DeepMind chair Demis Hassabis. TechCrunch published its report on the launch on September 17.

In one of the institute's inaugural essays, Hassabis proposes a U.S.-led frontier AI standards body to evaluate the most advanced AI models. The excerpt does not name any existing regulator, agency or law that the proposed body would replace or supplement, and it does not specify how the body would be funded, staffed or convened. It also does not state whether Google DeepMind would submit its own models for evaluation, or what would happen if it declined. The excerpt does not identify the authors of the other inaugural essays, or say how many were published.

The institute's three named directors connect it directly to Google DeepMind: Legg is a DeepMind co-founder and Hassabis is Google DeepMind's chair, while Manyika is identified as a Google executive. TechCrunch describes the launch as an effort to widen the AGI debate, and the only specific mechanism described in the excerpt is the standards body proposed in Hassabis's essay. The excerpt does not state whether the institute is a Google entity or an independent organisation. The excerpt does not include comment from Google, Google DeepMind, or any regulator or legislator on the proposal.