OpenAI bans Russian ChatGPT accounts used in covert misinformation campaign
The AI company said the incident "illustrates how influence actors can use AI as a supporting tool within a broader effort to manufacture authority."
Reasoning Overview
To meet the request, I first read the source article carefully, noting every concrete fact, figure, name, date, quote, and specific detail that must be retained exactly as reported. I then planned a rewrite that would:
1. Preserve all factual elements (OpenAI’s ban, the Russian‑origin accounts, the “sovereignty” index, the International Burke Institute, the platforms used, the February Rybar shutdown, the involvement of VPNs, and the outreach to Russian diplomatic channels).
2. Rephrase every sentence in original language, avoiding any verbatim copying or close paraphrase of the source.
3. Structure the piece with a lead paragraph followed by subheadings introduced by “
”, Each Subheading Limited to 3‑7 Words Drawn from the Article’s Own Terminology.
4. Keep the word count between 350 and 500 words, expanding where appropriate with general, publicly known context about OpenAI’s policy on Russian access and the broader pattern of AI‑enabled misinformation, without adding new specific data.
5. Use clear American English, active voice, and balanced attribution (e.g., “OpenAI said,” “the company stated,” “CNBC contacted…”) while avoiding any boilerplate or promotional language from the original publisher.
6. Omit any markdown formatting, bullet points, or decorative symbols, delivering plain‑text paragraphs separated by blank lines.
After drafting, I verified that each fact remained intact, that no sentence duplicated the source wording, and that the subheadings complied with the rule of using only terms already present in the article or the immediately following paragraph.
Rewritten Article
OpenAI Bans Russian‑Linked ChatGPT Accounts Over Hidden Influence Scheme
OpenAI announced Tuesday that it had disabled a cluster of ChatGPT accounts traced to Russia after uncovering a covert online influence operation that used the AI tool to generate and disseminate misinformation. The company said the investigation began with a review of AI‑produced social‑media posts, which then revealed a much larger coordinated effort.
Campaign Infrastructure
According to OpenAI’s blog post, the operation centered on a website that republished academic material with false attributions, a self‑styled “sovereignty” index that portrayed Russia in a positive light, and deliberate tactics to mask the operators’ Russian origins. Although OpenAI blocks access to its models from Russian IP addresses, the actors employed virtual‑private networks to circumvent those restrictions.
Prior Shutdowns and Connections
The latest takedown follows a February action in which OpenAI terminated ChatGPT accounts linked to Rybar, a pro‑Russia media outlet described by the United Kingdom as “partially coordinated by the Presidential Administration,” funded by the state corporation Rostec, and reportedly working with members of Russian intelligence services.
Content Generation Tactics
OpenAI reported that the operators entered prompts in Russian, directing the model to produce comments for platforms such as Substack, Telegram, X, Facebook and LinkedIn. They also instructed the system to strip any linguistic markers that might reveal a Russian origin. A significant portion of the fabricated material promoted the International Burke Institute (IBI), whose website hosts articles lifted from genuine academic sources, sometimes with invented attribution.
“sovereignty Index” and Reach
The influence campaign further introduced a “sovereignty index” that lauded Russia while disparaging Western nations. OpenAI assessed that the immediate audience reach was modest, but emphasized that the operation’s importance lay in the sophisticated infrastructure it built—an apparently credible institution complete with supposed experts, republished scholarship and a proprietary risk metric.
Broader Implications
OpenAI concluded that the episode demonstrates how influence actors can weaponize artificial intelligence to fabricate authority, conceal the provenance of favored narratives, and create scalable assets for future propaganda. The company also noted that the very use of AI tools can eventually expose such operations.
CNBC has reached out to the Russian embassy in London and Russia’s Foreign Ministry for comment. OpenAI’s decisive action underscores its commitment to preventing misuse of its technology and protecting the integrity of online discourse.
Source: CNBC · 2026-08-25