engblogs

Technical case study of AI-assisted deceptive employment campaigns that automate resume generation, fabricate remote-job personas, and evade security measures, with links to DPRK actors.

A technical case study of Operation Helgoland Bite, detailing how Russian-origin AI-generated German-language political content circulated via Telegram and X, linked to the Pravda network and Portal Kombat to influence the German discourse around Ukraine, NATO, and domestic politics, and the countermeasures leading to OpenAI bans.

Technical analysis of STORM-2035, an Iranian-linked recidivist influence campaign that used AI-generated multilingual comments to seed inauthentic chatter on X with minimal amplification.

An in-depth look at Operation 'High Five,' detailing how an AI-assisted effort generated bulk, short, partisan comments across TikTok and Facebook to influence Philippine politics, including content analysis, PR pitches, channel coordination, and the OpenAI-led detection and takedown.

Case study of detecting and attributing AI-generated German-language political disinformation across Telegram, X, and domain networks linked to Russia during the 2025 German election.

Technical case study of an AI-assisted task scam leveraging AI-powered translation across SMS, WhatsApp, Telegram and other platforms—likely originating in Cambodia—to recruit UK targets and extract funds via a ping, zing, sting workflow.

Case study of an OpenAI policy breach turned political-influence operation in the Philippines, detailing how a group used AI to analyze posts, generate bulk comments in English and Taglish, and craft PR pitches across TikTok and Facebook to sway public opinion around President Marcos and Vice-President Duterte.

A technical analysis of how a China-origin influence operation used AI-generated content, avatar/profile pictures, and posting-time optimization to polarize US political discourse across X and Bluesky, with implications for detection, attribution, and cross-platform amplification.

Case study of a China-origin influence operation leveraging AI-generated content to orchestrate multilingual, inauthentic social-media engagement and internal document generation across TikTok, X, Reddit, and other platforms.

Case study of an AI-driven influence operation, 'Uncle Spam', that generated polarized US political content across platforms, deployed synthetic personas and image-based branding, and used data-extraction techniques to profile users, with implications for platform security and attribution.

STORM-2035: Recidivist influence activity—an Iranian-linked operation using Persian prompts to generate multilingual influence comments posted by inauthentic X accounts.

A technical case study of STORM-2035 detailing Iranian-linked recidivist influence activity that used ChatGPT to generate multilingual, inauthentic X comments to sway political discourse while separating workstreams.