AkiraBot, a sophisticated Python framework, used OpenAI's GPT-4o-mini model to generate customized spam messages and targeted over 400,000 websites' contact forms and chat widgets to promote dubious SEO services, successfully spamming at least 80,000 websites since September 2024.
AkiraBot is a Python-based spam framework that has been active since September 2024, targeting small to medium-sized business websites built on platforms like Shopify, GoDaddy, Wix, and Squarespace. The bot uses OpenAI's GPT-4o-mini model with hardcoded API keys to generate customized spam messages that promote low-quality SEO services under the 'Akira' and 'ServiceWrap' brands. The system scrapes website content using BeautifulSoup and feeds it to the AI model with the prompt 'You are a helpful assistant that generates marketing messages' to create tailored outreach messages that appear legitimate and bypass spam filters. SentinelLABS analysis revealed that AkiraBot targeted more than 400,000 websites and successfully spammed at least 80,000 websites. The framework employs sophisticated evasion techniques including multiple CAPTCHA bypass mechanisms using services like Capsolver, FastCaptcha, and NextCaptcha, browser fingerprint spoofing through inject.js scripts, and proxy rotation via SmartProxy service. The bot targets both website contact forms and live chat widgets, including Reamaze integrations, operating through a GUI that allows concurrent multi-threaded attacks. OpenAI's security team collaborated with the investigation and disabled the associated API keys upon learning of the misuse.
Domain classification, causal taxonomy, severity scores, and national security assessments were LLM-classified and may contain errors.
Using AI systems to gain a personal advantage over others such as through cheating, fraud, scams, blackmail or targeted manipulation of beliefs or behavior. Examples include AI-facilitated plagiarism for research or education, impersonating a trusted or fake individual for illegitimate financial benefit, or creating humiliating or sexual imagery.
Human
Due to a decision or action made by humans
Intentional
Due to an expected outcome from pursuing a goal
Post-deployment
Occurring after the AI model has been trained and deployed