From TechTarget: 2024-05-31 15:05:27

OpenAI has detected and disrupted five influence operations globally, predominantly utilizing AI to generate text for social media posts and coding tasks. None of these operations were highly impactful on the Brookings Breakout Scale, with scores of only 1 or 2. These campaigns primarily spread political messaging across various platforms and languages.

Notable operations include Spamouflage from China, using AI for debugging code and posting content on multiple platforms. Bad Grammar from Russia targeted individuals on Telegram with political comments in multiple languages. Doppelganger, also from Russia, generated headlines and translated articles for social media. IUVM from Iran focused on generating and translating articles and headlines.

Stoic, a Tel Aviv company, operated Zero Zeno using OpenAI to create content for multiple websites. Meta recently took down accounts associated with Stoic, with only around 2,000 Instagram followers and 500 Facebook followers. OpenAI collaborates with industry partners to combat AI misuse and secure platforms.

To combat AI misuse, OpenAI collaborates with industry partners to design safer platforms and invests in technology to disrupt malicious actors. They leverage AI tools to combat abuses. Dark Reading has contacted OpenAI for further details on their efforts to combat and disrupt malicious actors.



Read more at TechTarget: OpenAI Disrupts 5 AI-Powered, State-Backed Influence Ops