New research from Barracuda has revealed Threat actors are now using OpenAI in impersonation campaigns targeting companies around the world.
The attack uses an email that mimics OpenAI and sends an ‘urgent message’ to victims recommending that they update their subscription payment details, all via their convenient direct link – a textbook example of a phishing technique.
The operation was far-reaching: one email was sent to more than 1,000 users. The first red flag was the sender’s email address, which did not match the official OpenAI domain (e.g. @openai.com). Instead, it was sent from info@mta.topmarinelogistics.com.
AI-powered
Worryingly, the email passed DKIM and SPF checks, meaning the email was sent from a server authorized to send emails on behalf of the domain. The language in the email is common in phishing attacks, pressuring the user to take immediate action and creating fear and urgency.
This is far from the only AI-related malicious campaign reported in recent months. Earlier in 2024, a Microsoft report found that 87% of UK organizations are more vulnerable to cyber attacks thanks to the increasing use of AI tools.
Not to mention the rise in deep-fake and persuasive AI voting fraud targeting businesses and consumers. Companies around the world have already lost millions to massive fake fraud, and almost half have been targeted by this type of scam at some point.
The introduction of machine learning algorithms that can discover and exploit software flaws means that AI is leading to a dramatic increase in attacks.
Despite this, research shows that 90% of cyber attacks still involve some element of human interaction, such as phishing attacks. So making sure everyone in your organization is trained to recognize the signs of an attack is the best protection for a business.