OpenAI has banned a cluster of accounts tied to an Israel-origin influence operation that used its models to mass-produce political content.
The company calls the operation Zero Zeno, and says it had not been previously reported. Accounts linked to the operation used OpenAI's tools to generate posts pushing anti-Hamas, anti-Qatar, and pro-Israel messaging, alongside separate anti-BJP and pro-Histadrut content aimed at audiences in India and Israel's labor movement. OpenAI removed the accounts as part of its ongoing work tracking misuse of its models for coordinated influence campaigns.
What stands out is the range. A single operation running pro-Israel, anti-Qatar, anti-BJP, and pro-Histadrut messaging at once looks less like an ad-hoc troll and more like a coordinated messaging shop working several political fronts in parallel. That's the pattern showing up across the state-linked influence operations AI companies have been disclosing lately: generative tools don't invent propaganda, they just make producing it at scale cheaper.
The real story isn't that an operation like this exists. It's that AI has lowered the cost of running five of them at once.