AI leadership? Trump posts deepfakes of Dems calling themselves “woke pieces of s–t.” (arstechnica.com)

President Trump posted a 35-second AI-generated deepfake to Truth Social showing synthetic versions of Senate Majority Leader Chuck Schumer and Rep. Hakeem Jeffries trading crude insults while mariachi music plays and one is adorned with a CGI sombrero and fake mustache. The clip uses both audio and video manipulation—voice cloning/faux speech and facial/overlay CGI—to create a realistic-seeming but fabricated interaction aimed squarely at political opponents amid budget negotiations. The episode matters because it normalizes high-profile misuse of generative media by senior officials, accelerating a decline in trust of recorded evidence and complicating fact-finding in politics. Technically, it underscores how current deepfake toolchains can produce short, persuasive clips that evade casual detection, heightening the need for provenance, cryptographic watermarks, robust forensic detectors, and platform policies to label or remove synthetic content. For the AI/ML community, this is a reminder that model capability advances must be paired with detection research, deployment safeguards, and governance mechanisms—otherwise synthetic media will increasingly be weaponized to mislead, inflame, and derail democratic processes.