Groundbreaking OpenAI products such as Whisper and ChatGPT have become immensely popular. Demonstrations of ChatGPT and other artificial intelligence applications presented several possible risks, including misinformation, disinformation, and various forms of malicious exploitation. Shortly after ChatGPT’s launch, cybercriminals developed phishing scams to capitalize on its popularity. The NJCCIC received reports of a newly created YouTube channel, “OpenAI Live,” attempting to scam users into donating cryptocurrency or directing them to malicious websites. In one example, the cybercriminals created a live stream YouTube fundraiser, broadcasting a previously recorded interview, and contained potentially malicious links in the comments. One pinned comment included the recently created URL, openai[.]gift, with a serving IP address of 45.9.74[.]100, which was identified as malicious by VirusTotal. This YouTube channel has since been restricted pending further review; however, threat actors continue to create new misleading channels and associated URLs. Other phishing scams and fake ChatGPT application downloads and extensions were discovered attempting to steal funds and personal information or deliver malware.

The NJCCIC recommends users refrain from responding to unsolicited communications, such as emails and social media feeds or messages, and avoid clicking links or opening attachments from unknown senders. Additionally, users are advised to avoid downloading OpenAI applications from unofficial sources and, instead, navigate directly to the official OpenAI webpage.

For any further questions, contact us here at Cyber Command.