Facebook Users Targeted by AI‑Driven Phishing Scam

On August 10 2026, Rossen Reports posted an urgent alert on its Facebook page warning of a new phishing scheme targeting platform users. The post described fraudulent websites that ask for personal information or direct users to click links leading to fake login pages. ¹

Investigators say attackers are using large language models (LLMs) such as those powering chatbots to produce highly personalized emails and texts with little prompting. In a controlled experiment last month, researchers fed these bots limited public data scraped from previous breaches and achieved persuasive content resembling human communication patterns. This shift lowers the skill floor for creating convincing phishing material and makes even less technically sophisticated users vulnerable. ²

The scam appears focused on elderly demographics living mainly in U.S. west‑coast states—California, Oregon, Washington—where internet usage is high but awareness about evolving cyber threats remains low. Attackers employ URL shorteners to mask malicious sites and exploit user complacency when entering credentials into untrusted pages after clicking unsolicited communications.

FTC guidance advises users to enable multi‑factor authentication whenever possible, remain cautious of unexpected emails/texts asking them to click links or open attachments, and verify URLs before submitting any sensitive information. These steps can help mitigate risk from AI‑generated phishing attempts. ³

Attackers craft messages that mimic official communications from Facebook or trusted contacts. By leveraging LLMs they can tailor language to match a user’s previous interactions and known interests, making detection harder. The resulting text often contains subtle linguistic cues that resemble genuine social media posts, thereby reducing suspicion among recipients who might otherwise flag generic spam. ²

Once a victim clicks on the malicious link, they are redirected to a site that looks identical to Facebook's login page. The site requests credentials which are then captured by attackers. Because the page uses HTTPS and displays familiar logos, many users fail to notice subtle differences such as mismatched URLs or lack of security indicators like “⁴ versus shortened domains. Users should always check the address bar for exact domain spelling before submitting credentials. ¹

Experts warn that older adults may be more susceptible due to lower digital literacy and higher trust in online platforms.

Overall, the August 10 alert highlights a new wave of phishing that combines advanced language models with traditional social engineering tactics. Targeting older adults and using fake login pages, attackers raise detection difficulty while keeping execution simple. Enabling multi‑factor authentication, verifying URLs, and remaining cautious of unexpected messages can help protect against this evolving threat.