AI Social Engineering: Microsoft Warns of Targeted Brand Bait Scams

(Wait, let’s keep the numbered list since the rules encourage lists).

Let’s write ChatGPT & Claude section as:
`

Beyond malvertising, email delivery vectors have been heavily weaponized. Microsoft’s threat advisory detailed a massive ChatGPT-themed phishing campaign that dispatched up to 100,000 emails daily, masquerading as urgent billing alerts for ChatGPT Plus. To bypass automated email security filters, the phishing flow routed victims through a complex, multi-stage redirect chain using legitimate infrastructure, such as Amazon tracking domains, before landing them on compromised domains designed to harvest credit card data.

Simultaneously, cybercriminals targeted users of Anthropic’s Claude with fake account suspension appeals. In this campaign, attackers sent urgent notifications claiming that the user’s account had violated service policies and was scheduled for immediate termination. The phishing flow functioned as follows:

1. The victim received an email containing a link to a fake policy PDF or a direct appeal page.
2. To access the appeal portal, the user was prompted to authenticate through an Adversary-in-the-Middle (AiTM) phishing setup.
3. The AiTM framework proxied the authentication request to the legitimate identity provider, allowing the user to complete their standard multi-factor authentication (MFA) flow.
4. The proxy intercepted the resulting session cookies, allowing the attackers to hijack the active session entirely and bypass MFA controls.

` (215 words) – Saved