AI-Generated Phishing

AI-generated phishing uses generative AI to write convincing scam emails, texts and fake websites quickly, in any language and personalized to each target, removing the spelling mistakes and awkward wording that used to give scams away.

BeginnerCategory: AI and emerging threats

Also called: AI phishing, LLM phishing, generative AI phishing

What is AI-generated phishing?

Phishing has always relied on persuasive writing. Generative AI tools write persuasively in seconds, in any language, in any tone. AI-generated phishing is the use of those tools to produce scam messages and fake websites.

The FBI warned in December 2024 that criminals use AI-generated text to make social engineering and spear phishing more believable, to send messages faster to a wider audience, and to fix the grammar and spelling errors that used to expose foreign scammers targeting US victims.

How it works

  • Flawless writing: no typos, no awkward phrasing, natural local idioms.
  • Personalization at scale: AI can research a target's LinkedIn profile, company news and social media and write a tailored message for each of thousands of people, work that once took a skilled scammer hours per target.
  • Convincing impersonation: matching the writing style of a real executive or vendor from emails stolen in an earlier breach.
  • Instant fake websites and documents: cloned login pages, fake invoices and convincing landing pages.
  • Chatbots that hold a conversation, replying naturally when a victim asks questions.
  • Combined with deepfakes: a phishing email followed by a cloned voice call to confirm it.

Mainstream AI tools try to refuse obvious phishing requests, but criminals get around safeguards with clever prompting or use unrestricted models.

Why old advice fails

"Look for bad spelling and grammar" no longer works. Training needs to focus on the request, not the writing: urgency, secrecy, payment changes, credential requests and unusual channels.

How to defend against it

  1. Verify requests independently, especially for payments, bank changes and password resets, by calling a known number.
  2. Use phishing-resistant MFA, such as passkeys and security keys, so stolen passwords and codes aren't enough. CISA recommends it as the strongest protection.
  3. Use modern email security that analyzes behavior and sender reputation, not just keywords, and set up DMARC for your domain.
  4. Update [security awareness training](/glossary/security-awareness-training/) with realistic AI-written examples.
  5. Make reporting easy with a report-phishing button and a no-blame culture.

See our guide to phishing attacks.

Related terms

Sources

  1. Criminals Use Generative Artificial Intelligence to Facilitate Financial Fraud (FBI Internet Crime Complaint Center (IC3))
  2. Implementing Phishing-Resistant MFA (CISA)

Drafted with AI assistance, then edited and fact-checked by a person. How we use AI.