A sophisticated phishing campaign is exploiting ChatGPT branding via fake billing notices to harvest user credentials and compromise OpenAI accounts.
Security researchers have uncovered a new phishing campaign utilizing fraudulent ChatGPT billing notifications to direct unsuspecting users toward credential-harvesting pages. The attack specifically targets both personal and enterprise accounts by impersonating standard subscription renewal notices.
Josh Varden of Cofense’s Phishing Defense Center analyzed the malicious emails and discovered that the messages arrive under the subject line “Urgent: Update Your Payment Method to Avoid Service Interruption.” The communications feature official ChatGPT branding, a simulated final-notice warning, and a fabricated outstanding balance of $23.80, pressuring victims to act within a 48-hour window.
Technical analysis reveals that the emails originate from the suspicious sender address support@9527db6e1a[.]nxcli[.]io rather than legitimate OpenAI infrastructure. When recipients click the embedded “Update Payment Information” button, the link routes through a Google API redirect at notifications[.]googleapis[.]com before landing on the attacker-controlled server hosted on the nxcli[.]io domain.
The resulting phishing landing page replicates the authentic OpenAI sign-in interface, featuring the ChatGPT logo and a standard username and password prompt. Credentials entered into these fields are immediately captured by attackers, while victims are subsequently redirected to a generic error page.
Security teams are advised to monitor mail logs for specific indicators of compromise, including the identified Google API redirect mechanism alongside the login.php and key.php paths hosted on the nxcli[.]io domain. Enterprise IT administrators should reinforce user awareness training regarding verified authentication pathways, reminding employees to always verify that the browser address bar displays auth.openai.com before submitting sensitive credentials.