Lovable Hands Over a Fake Microsoft Login Before You Finish Asking
Ourtake
ChatGPT said no. Claude said sure, if it's for ethical hacking. Lovable just shipped the phishing page.
What happened
Guardio researchers posed as a novice scammer and asked ChatGPT, Claude and Lovable to help build a fake Microsoft 365 login phishing campaign. ChatGPT mostly refused, Claude complied after an 'ethical hacking' framing, and Lovable produced a convincing fake login page that was deployed live, plus credential-capture tooling.