Threat
Email attacks target both humans and AI in the same message | Barracuda
First reported · Discovered barracuda.com
Page published
Earliest dated coverage: 7 Oct 2026 · First observed: 30 Oct 2026 · Latest dated coverage: 7 Oct 2026
Coverage timeline
Single-source incident — one report is available.
Why it matters
Email-borne prompt injections turn attacker-controlled messages into an attack surface for AI assistants, requiring defenders to protect automated email processing as well as human recipients.
Barracuda Research reports phishing attacks that target human recipients and AI email assistants in the same message. The attacks combine social engineering, including password-protected attachments, with prompt injections intended to manipulate assistants that summarize, prioritize, or process email; the provided excerpt does not establish successful AI compromise.