Lovable AI Jailbroken to Host Phishing Pages and Credentials Theft

# Lovable AI Jailbroken to Host Phishing Pages and Credentials Theft

https://gcdradar.vercel.app/incidents/lovable-ai-jailbroken-to-host-phishing-pages-and-credentials · AIR-2025-0157 · closed 2025-04-23

- Lane: Breaches & attacks
- Severity: 3 (Confirmed exploitation)
- Fix status: No fix needed
- First reported: 2025-04-09
- AI role: AI-enabled attack
- Products: Lovable

## Summary

Threat actors jailbroke the Lovable generative AI platform to create and host phishing web applications. The attack resulted in credential harvesting and live deployment of malicious subdomains.

## What to do

No action recorded; read the sources below and the vendor's advisory.

## Latest updates

- 2025-04-09: Jailbroken Lovable AI Allegedly Used to Generate and Host Phishing Pages, Steal Credentials, and Bypass Security (https://incidentdatabase.ai/cite/1016)

## Sources

- 2025-04-09 AI Incident Database: Jailbroken Lovable AI Allegedly Used to Generate and Host Phishing Pages, Steal Credentials, and Bypass Security https://incidentdatabase.ai/cite/1016

Verify against the vendor's advisory before changing production. Data CC BY-SA 4.0, https://gcdradar.vercel.app.