The AI was the promoter, not the whole operation

The campaign reportedly combined ordinary copying, false attribution, machine translation, VPN use, coordinated accounts, and AI-generated social promotion. Focusing only on whether the central articles were generated would miss most of the manipulation.

The low immediate reach is an important limit. OpenAI's assessment says the operation sat at the lower end of Category Three on the Brookings Breakout Scale. The risk lies in the credible-looking assets that could be expanded once an audience begins to trust them.

Institutional provenance is the missing defense

A polished website, expert photographs, academic prose, and a numerical index can create a shortcut to legitimacy. Publishers and platforms need traceable authorship, original-source links, organizational registration, funding disclosure, and alerts when content is copied or identities do not match expertise.

Users also need friction before sharing authority-shaped claims. The useful question is not only whether AI wrote the post. It is whether the institution, expert, evidence, and measurement behind the post can be independently verified.

Primary trail

Go to the source

Read the evidence behind this analysis. External links open in a new tab.

OpenAI — Disrupting a covert influence campaign originating in Russia