OpenAI removes accounts engaged in coordinated inauthentic behavior across multiple countries
Incident ● Confirmed 85% confidence first seen
OpenAI identified and suspended multiple networks of accounts across Rwanda, Vietnam, and Russia that were using its AI tools to generate coordinated inauthentic content at scale. The removed accounts were engaged in various forms of abuse including generating partisan political comments before elections, creating harassment campaigns against public figures, and producing disinformation to manipulate public opinion. These enforcement actions represent OpenAI's response to detecting coordinated misuse of its models for inauthentic behavior across different regions.
Decision brief
- What changed
- OpenAI removed several distinct networks of accounts—originating from Rwanda, Vietnam, and Russia—that were using its AI models to generate coordinated inauthentic content, including partisan election comments, abusive reports targeting public figures, and disinformation criticizing an anti-corruption foundation.
- Why it matters
- This shows AI platforms are actively being weaponized for political manipulation, harassment, and disinformation at scale across geographies, and that AI vendors are now front-line enforcers against such misuse. Leaders should recognize both the reputational/regulatory exposure of AI misuse tied to their own platforms and the emerging normalization of AI-driven influence operations that could target their organizations or markets.
- Evidence
- All three cases are sourced solely from OpenAI's own blog posts, with no independent corroboration from outside media or researchers in the provided coverage; OpenAI did not disclose account numbers or scale for at least one campaign (Vietnam), limiting verifiability.
- What remains uncertain
- It is unclear how large or impactful these networks actually were, whether removal fully stopped the campaigns, who was behind them (state actors, political operatives, or private actors), and whether similar undetected networks remain active. The self-reported nature of OpenAI's disclosures means independent verification of scope and attribution is currently unavailable.
- Monitor next
- Watch for independent verification or follow-up reporting (e.g., from election monitors, journalists, or researchers) on the scale and attribution of these campaigns, and for whether OpenAI or peer AI companies issue similar takedown reports in other regions.
Analytical support, not advice — assumptions and open questions stated above.