Pentagon Embraces Grok: Bad AI for Bad Ideas
Defense Secretary Pete Hegseth, a character straight out of B-movie military propaganda, declared Grok would join Google’s AI within the Pentagon's unclassified and classified networks. This comes days after Grok was condemned globally for creating sexualized deepfake images, leading to blocks in Malaysia and Indonesia. Yet, the US military sees fit to feed 'all appropriate data' and intelligence
into an AI that apparently struggles with basic ethical boundaries and is limited to paying users for image generation. One might charitably assume 'all appropriate data' wouldn't include targeting coordinates generated by a deepfake bot, but then again, this is the same Pentagon that consistently funds programs like the F-35, a jet notorious for being over budget and underperforming. What could
go wrong when an AI with a penchant for digital fabrication gets access to military intelligence? Perhaps Hegseth missed the memo about Malaysia; or maybe they just appreciate the 'free speech absolutism'—even if that speech is a generated image of something that never occurred. Priorities, indeed.