OpenAI has banned a network of ChatGPT accounts that were leveraging its models for social engineering, surveillance-themed research, and influence operations targeting critics. The operation, dubbed "VAGue Focus," originated from mainland China and primarily used AI to generate social media posts, analyze datasets, and translate communications. The actors created personas for entities like "Focus Lens News," "Bright Wave Media Europe," and "Visionary Advisory Group" (VAG), posing as journalists and geopolitical analysts to collect intelligence.

The operation utilized ChatGPT for four main workstreams: generating social media content and biographies for covert influence, polishing correspondence to a US Senator, inquiring about cyber attack tools with basic questions, and translating Chinese messages designed to extract information from unknown individuals. Some of these messages were posted as replies to journalists and researchers on X, while others resembled direct messages.

The primary objective appeared to be intelligence gathering by posing as professionals in Europe or Turkey. "Visionary Advisory Group" allegedly offered $2,000 per hour for interviews regarding US economic and financial policies and sought classified documents. While the operation used AI for data scraping and analysis, its public-facing impact was limited, with the Focus Lens News X account, despite having 17,000 followers, showing signs of being compromised and repurposed. The overall activity was assessed as low-impact.