KUALA LUMPUR, June 17 – Palo Alto Networks, the global leader in cybersecurity, has unveiled groundbreaking research conducted by its threat intelligence team, Unit 42, shedding light on the rising wave of scams exploiting the growing popularity of generative AI, specifically targeting users of ChatGPT. The study reveals the multifaceted tactics employed by scammers to deceive individuals into sharing sensitive information or unwittingly installing malicious software. Concrete instances and case studies are provided to illustrate these methods.
Unit 42 delved into numerous phishing URLs masquerading as the official OpenAI website. Perpetrators of such schemes create counterfeit websites that closely mimic the genuine ChatGPT platform, aiming to deceive users into downloading harmful software or disclosing private and confidential information. Despite OpenAI offering a free version of ChatGPT, scammers often mislead victims into visiting fraudulent websites and paying for supposed services.
Key findings from the report include, the AI ChatGPT extension can inject a background script into victims’ browsers, featuring highly obfuscated JavaScript. This script leverages the Facebook API to pilfer victims’ account details, potentially gaining further unauthorized access. Unit 42 observed a staggering 910% increase in monthly domain registrations related to ChatGPT between November 2022 and April 2023. The Palo Alto Networks Advanced URL Filtering system detected over 100 malicious URLs linked to ChatGPT on a daily basis.
During the same period, the team witnessed a remarkable 18,000% growth in squatting domains based on DNS security logs. Multiple phishing URLs were identified, attempting to impersonate official OpenAI sites. Scammers meticulously construct fake websites that closely resemble the legitimate ChatGPT platform, coercing users into downloading malware or revealing sensitive information. Despite the availability of a free ChatGPT version from OpenAI, scammers mislead victims into visiting fraudulent websites under the pretense of requiring payment for these services.
