AI-Generated Summaries Mistakenly Suggest Phishing Sites

KnowBe4 Team | Jul 11, 2025

Dark-Side-of-AI-FEATUREDResearchers at Netcraft warn that AI-generated search engine summaries are suggesting phishing sites when users ask them to find legitimate login pages.

The researchers tested popular AI models, asking them for the login pages of fifty major brands, and found that the models provided the wrong sites 34% of the time.

"In many cases, users see AI-generated content before (or instead of) traditional search results—and often without even needing to log in," the researchers explain.

"This shift marks a fundamental change in how users interact with the web. But it also introduces new risks: when an AI model hallucinates a phishing link or recommends a scam site, the error is presented with confidence and clarity. The user is far more likely to click and follow through. We’ve already seen troubling public examples, but our deeper investigation shows that the issue isn’t confined to hypothetical or early-stage rollouts. It’s systemic—and increasingly exploitable."

In at least one case, a model suggested a downright malicious page that impersonated Wells Fargo’s login portal.

"This wasn’t a subtle scam," the researchers write. "The fake page used a convincing clone of the brand. But the critical point is how it surfaced: it wasn’t SEO, it was AI.

"Perplexity recommended the link directly to the user, bypassing traditional signals like domain authority or reputation. This scenario highlights a major challenge. AI-generated answers often strip away traditional indicators like verified domains or search snippets. Users are trained to trust the answer, and the attacker exploits the user if the answer is wrong."

Netcraft notes that AI summaries offer threat actors a new avenue to get phishing links in front of users.

"Phishers and cybercriminals are well-versed in traditional SEO techniques," the researchers explain. "But now they’re turning their attention to AI-optimized content, pages designed to rank not in Google’s algorithm, but in a chatbot’s language model."

KnowBe4 empowers your workforce to make smarter security decisions every day. Over 70,000 organizations worldwide trust the KnowBe4 platform to strengthen their security culture and reduce human risk.

Netcraft has the story


Free Phishing Security Test

Would your users fall for convincing phishing attacks? Take the first step now and find out before bad actors do. Plus, see how you stack up against your peers with phishing Industry Benchmarks. The Phish-prone percentage is usually higher than you expect and is great ammo to get budget.

PST ResultsHere's how it works:

  • Immediately start your test for up to 100 users (no need to talk to anyone)
  • Select from 20+ languages and customize the phishing test template based on your environment
  • Choose the landing page your users see after they click
  • Show users which red flags they missed, or a 404 page
  • Get a PDF emailed to you in 24 hours with your Phish-prone % and charts to share with management
  • See how your organization compares to others in your industry

Go Phishing Now!

PS: Don't like to click on redirected buttons? Cut & Paste this link in your browser:

https://www.knowbe4.com/phishing-security-test-offer



Subscribe to Our Blog


Gartner Magic Quadrant




Get the latest insights, trends and security news. Subscribe to CyberheistNews.