ChatGPT fans need to be mindful of scammers and implement a “defensive mindset” when interacting with the bot. Cybercriminals have already started unleashing innovative ways to trick users into downloading malware and disclosing confidential data.
Per Unit 42—Palo Alto Network’s threat intelligence apparatus—domain-based references of ChatGPT increased 910 percent between November 2022 and April 2023. For the same period, researchers also noted a 17,818 percent surge in squatting domains from DNS Security, as well as 118 daily detections of ChatGPT-related malicious URLs.
The sharp rises in these domains show that scammers are out to deceive ChatGPT users into visiting fake chatbot sites or malicious websites, so they can obtain their personal data. Mostly, fraudsters create sites that look similar to the official OpenAI ChatGPT page then dupe users into providing their credit card and email details by using technology and crypto-fraud scams in association with Elon Musk’s name and image.
Palo Alto also observed multiple cases of mischievous registrants utilizing openai and chatgpt related domains such as openai.us and chat.gpt.jobs. Though none of this domains hosted any malicious contents by early this month, there is a greater chance of them being exploited in the future as they’re not under OpenAI or legitimate domain administration firms’ oversight.
Starting from November, a constant rise in squatting and other related domains has been seen. The upward trend peaked in February after Microsoft—the primary investor funding OpenAI’s technologies like GPT-4, ChatGPT and Dall-E included functionalities into the Bing search engine.
The report also included small portions of inferior ChatGPT-based AI bots, some replete with huge language models and others claiming to offer access to the OpenAI API. These bots are bearers of security threats, especially in regions where ChatGPT isn’t accessible. In one case, researchers downloaded a malicious Chrome extension in the name of “AI ChatGPT”, which sneaked in and obtained confidential data such as email addresses and credit card information, as well as backdoor access with super admin control.
To keep dodgy websites and copycats in check, it is advised that ChatGPT users be cautious when clicking suspicious emails or links, and only use the OpenAI website for ChatGPT access. Furthermore, Guardio, an antivirus provider, maintained that everyone should keep a close eye out for extensions and browser hijacking as well.
Palo Alto Networks is an American global cybersecurity firm that provides security solutions for cloud, networks, and endpoints. It has over 10,000 employees around the world and serves over 83 countries. It creates and offers technology to keep computer networks of both public and private sectors safe.
Unit42 is the threat intelligence unit under Palo Alto Network. It is an advanced research team dedicated to the discovery and reporting of the latest cyber threats. It releases reports on the latest cybersecurity research, malware trends, and more. Lucas Huang, Peng Peng, and Zhanhao Chen are the three researchers mentioned in the article. They are part of the threat intelligence unit and have been researching various ChatGPT-related malicious activities.
