Meta, Google Hand Out Deepfakes, Censor Your Speech

Meta and Google are integrating AI-powered image and video generators into Instagram, WhatsApp, and Google Photos, enabling users to create and edit photorealistic content with simple text prompts. The tools, including Meta's Muse Image and Google's Video Remix, raise concerns about deepfake proliferation while both companies maintain strict content moderation policies. Critics argue the same platforms democratizing synthetic media also control speech through bans and shadow-banning.

Meta and Google are racing to put AI-powered image and video generators directly into the apps billions of Americans use every day, handing ordinary people the ability to fabricate photorealistic content — while both companies maintain tight control over who gets to speak on their platforms. Meta is rolling out "Muse Image" for Instagram and WhatsApp, which lets users create and edit realistic photos using simple text prompts, including changing lighting, generating entirely new images, or restoring old family photos, according to WEAU. A video version, "Muse Video," is expected in the coming months. The push is part of Meta's billions-in spending spree to compete with OpenAI and Google in the AI arms race. Not to be outdone, Google is launching "Video Remix" inside Google Photos, powered by its new Gemini Omni model. The feature lets users apply AI-driven edits to saved videos — swapping backgrounds, overlaying watercolor filters, or relighting dark scenes — all in seconds and without any editing skill. Engadget reports that Gemini Omni also allows users to insert themselves into videos via a digital avatar. Google says the tool is the successor to its Veo 3.1 video generator and better understands physical forces like gravity, making generated scenes more convincing. Translation: the tools for manufacturing convincing fake footage are being democratized at breakneck speed. Google is at least acknowledging the deepfake problem on paper. Engadget notes that Gemini Omni's avatar feature is watermarked with Google's SynthID tool — a self-regulatory gesture from the same company that has spent years manipulating search results and throttling reach for content it deems problematic. But SynthID is Google's own system, with no independent verification, and it only covers the avatar feature. The broader editing capabilities — background swaps, relighting, style filters — carry no such safeguards in any of the reporting. 9to5Google framed Video Remix as a fun way to create "share-worthy moments" without "professional skills or hours of editing." Engadget pitched it as saving users from "hours of Premiere Pro tutorials." Neither outlet lingered on the obvious: the same tools that let you add a watercolor filter also let you fabricate a scene that never happened, and the barrier to doing so just collapsed to zero. Meanwhile, both companies continue to enforce speech codes across their platforms, banning and shadow-banning accounts that challenge establishment narratives. The same platforms that will host millions of AI-generated images and videos — potentially indistinguishable from real ones — still reserve the right to silence you for stating biological facts or questioning official stories. Video Remix is rolling out now to Google AI Plus, Pro, and Ultra subscribers in the U.S. and select countries, according to both 9to5Google and Engadget. Meta's Muse Image launch timeline wasn't specified beyond a general rollout. The question isn't whether Americans will use these tools creatively. They will. The question is who gets to decide what's real when the companies fabricating reality also control who's allowed to talk about it.