When Every Video Might Be Fake, Witnesses Ask You to Stop Sharing the Ones That Are

An account on X posted a direct plea this week, addressed to someone sharing a video from the front: "Please don't spread misinformation, this video is AI, we have suffered enough and we have more real horrendous footage — we will lose our honesty because of this fake stuff." The post got over a hundred likes and was retweeted dozens of times, not because it was spectacular but because it was plainly desperate. The author wasn't a fact-checker or a journalist. They were someone trying to protect the credibility of their own testimony against a tide of AI-generated fabrications that nobody asked for and almost nobody can stop.

This is where the generative AI misinformation problem gets genuinely difficult to talk about. The formal discourse — think tanks publishing reports on Iranian TikTok campaigns, Pentagon studies on cognitive domain operations, the Foundation for Defense of Democracies worrying about deepfakes on front lines — treats the problem as one of adversarial state actors manipulating passive audiences. That framing has real value. But it misses the texture of what's actually happening in communities closest to the conflict. On X, a fan account with 36 retweets is mobilizing followers to report a TikTok user for uploading AI-modified content of their favorite idols. A different account is coordinating a block-and-report campaign against a profile that created a deepfake of a person named Dani in a McDonald's uniform to humiliate her publicly. These aren't geopolitical operations. They're targeted harassment campaigns, parasocial violations, and the weaponization of cheap image tools against ordinary people — and the communities fighting back are doing so with the only tools available to them: mass reporting and mutual aid.

Meanwhile, on Bluesky, a post with over four hundred likes made a blunter argument about institutional failure: Google has become perhaps the largest single source of misinformation in the world, the author wrote, pointing to an AI-generated result and calling it completely fabricated. It's a charge that lands differently than it would have two years ago. The Gaza footage plea and the Google accusation are separated by geography and context, but they're pointing at the same underlying collapse: when AI can generate plausible content faster than any verification infrastructure can process it, the burden of proof shifts onto victims and witnesses. They become responsible for authenticating their own reality to an audience that has learned, reasonably, to distrust everything.

Deepfake detection is suddenly appearing in conversations where it barely registered a week ago, which suggests people are starting to reach for technical solutions. But detection tools solve a different problem than the one witnesses are describing. The person begging you not to share the AI video doesn't need a classifier — they need the sharer to pause before clicking repost. That pause is a social and epistemic problem, not a technical one, and no model trained on synthetic media artifacts is going to manufacture it. The communities doing the hardest work here — coordinating reports, authenticating footage, publicly naming bad actors — are improvising in the absence of platform infrastructure that should already exist. They're not waiting for the think tanks to catch up.