Send spam to 80,000 domains

Advanced Akirabot robot, developed in Python programming language, has been able to produce personalized spam messages using the Openai API and send them to websites via online call and chat forms.

According to Sentinellabs, the robot has shortly targeted more than 80,000 domains and failed in only 11,000.

The use of artificial intelligence in spam production

Unlike traditional spam sending tools that use duplicate templates, Akirabot produces unique content to each website. Using the OPI OPI and the BeautifulSoup tool, the robot extracts and designs the content of the websites and designs messages that pass spam filters. The main purpose of the campaign was to promote suspicious SEO services with Akira and Servicewrap brands.

Akirabot uses sophisticated techniques to circumvent Capchay systems such as “Recaptcha” and “HcAPTCHA”. With tools such as Seleenium and Inject.js, this robot can simulate human user behavior and even manipulate system features such as fonts and memory.

Also, using proxy services such as SmartProxy, it changes IP addresses to escape IP -based identification and restrictions.

The Sentinelabs report shows that the robot initially started operating with Shopify -based sites, then expanded to platforms such as “Godaddy”, “WIX” and “Squarepace”. The robot’s graphic interface allows operators to monitor results and adjust the same goals.

Akirabot attack is a worrying example of artificial intelligence abuse in spam campaigns. Although Openai quickly disabled the robot’s API key, experts warning that operators in the campaign are likely to improve their methods. These attacks represent the increasing complexity of artificial intelligence -based cyber threats that need to strengthen the security measures of website hosts.