Operation "Doppelganger": Russian influence activity targeting Ukraine

OpenAI's latest threat report reads less like a tech update and more like a counterintelligence memo. The company confirmed it shut down accounts linked to Doppelganger, a Russian-origin influence operation that's been sowing anti-Ukraine sentiment online since at least 2022. What's new here isn't the operation itself — researchers have tracked Doppelganger for years — but the tooling behind it.

According to OpenAI, the banned accounts were using its models to mass-produce social media comments, generate translations, and write website copy, all aimed at undermining support for Ukraine across multiple languages. This is the grunt work of propaganda: the repetitive, volume-heavy content that used to require rooms full of paid trolls typing in broken English. Now a language model does it faster, cheaper, and in fluent German, French, or Polish without breaking a sweat.

The operation's playbook hasn't fundamentally changed — flood comment sections, mimic legitimate news outlets, push narratives that fracture Western unity on Ukraine. But the scale problem that used to limit these campaigns is dissolving. A single operator with API access can now do what once took a small team, and do it in a dozen languages simultaneously.

OpenAI didn't share exact numbers on how many accounts were caught or how much content got generated before the bans landed. That's a pattern with these disclosures: enough detail to prove the threat is real, not enough to map the full scope. Still, the acknowledgment itself matters. It's another data point in a growing list of state-linked actors — Russian, Chinese, Iranian — caught trying to weaponize commercial AI tools for influence operations, and it confirms Ukraine remains squarely in the crosshairs.