OpenAI removes accounts across multiple coordinated inauthentic behavior campaigns using AI for political influence operations
Incident ● Confirmed 92% confidence first seen
OpenAI disclosed enforcement actions against five separate coordinated influence operations—originating from the Philippines, China, Iran, and Russia—that were using its AI tools to generate inauthentic political content and disinformation across multiple countries and platforms. The operations targeted various geopolitical objectives including US polarization, elections in multiple countries, and criticism of specific political figures. OpenAI's enforcement disrupted these campaigns by removing the associated accounts and preventing further scaling of AI-generated political influence content.
Decision brief
- What changed
- OpenAI disclosed enforcement actions removing accounts across five separate coordinated inauthentic behavior campaigns—linked to the Philippines, China (two operations), Iran, and Russia—that used its AI tools to generate political influence content targeting the US, UK, Ireland, Venezuela, Germany, Taiwan, and the Philippines.
- Why it matters
- This confirms state- and non-state-linked actors are actively operationalizing generative AI for cross-border political influence and disinformation at scale, raising platform, regulatory, and reputational exposure for AI providers and any company whose tools could be similarly misused. It signals that trust & safety enforcement against AI-enabled influence operations is now a recurring operational function, not a one-off event, with implications for compliance, PR, and platform policy teams.
- Evidence
- All information comes from five OpenAI blog posts self-reporting its own enforcement actions; there is no independent third-party verification, journalistic investigation, or corroboration from external researchers included in this coverage.
- What remains uncertain
- Scale metrics (number of accounts, reach, engagement, actual influence achieved) are not disclosed, so the real-world impact of these operations before takedown is unknown; attribution to specific state actors (China, Iran, Russia) relies solely on OpenAI's own assessment without independent confirmation.
- Monitor next
- Watch for independent verification or follow-up reporting from disinformation researchers (e.g., Stanford Internet Observatory, Graphika) or other AI platforms disclosing similar coordinated campaigns, which would indicate this is an industry-wide pattern rather than an OpenAI-specific finding.
Analytical support, not advice — assumptions and open questions stated above.