Urgent Security Concern with ChatbotBuilder AI

You decide.

[cache cleared. New support chat opened]

Me: “Security.”

Chatbot: “Is this (my full first and last name)?”

Either somebody needs a new job or somethin’ up - timeline of events
• Created an account on ChatbotBuilder AI a few hours ago.
• Received only a welcome email, no other communications.
• Began configuring chatbot, but noticed modifications occurring without my input.
• Support chat behaved strangely, switching between AI-generated responses and what seemed like live manual edits.
• Support was aware of my screenshots in real time, commenting on actions I was taking.
• Attempted to email their security and legal contacts—all emails bounced back as nonexistent.
• Started a fresh support chat session, reported the issue, and submitted a security ticket.
• No email confirmation of the ticket, despite the system acknowledging it was created.

:rotating_light: Tomfoolery :
• Unauthorized chatbot modifications without my input.
• AI and human-like inconsistencies in support chat, suggesting possible live intervention.
• Support aware of my screenshots in real time, indicating potential screen tracking or surveillance.
• Security/legal emails bouncing, meaning no way to verify legitimacy or escalate concerns externally.
• No email confirmation for security ticket, making the reporting system questionable.
• Support chat not following standard AI behavior, responding in a way that seemed altered on the fly.
• System asked for personal information (billing email) before addressing concerns, an odd request in a security situation.

OpenAI integrated. Thoughts?

1 Like

This is not a product associated with OpenAI. They might be using the OpenAI API but OpenAI is not associated with them.

Cheers! :hugs:

1 Like

Good call out for those who may be unaware what this means for OpenAI :clinking_glasses:.

ChatBotBuilder.ai ≠ OpenAi, BUT;

If compromised as a wrapper, user inputs, API calls, and responses could be intercepted or manipulated.

Telemetry & authentication anomalies suggest backend interference.

:point_down:t2::point_down:t2::point_down:t2:
Breach Indicators & Forensic Analysis 70.4%

AI anomalies & unauthorized LLM modifications
Covert telemetry exfiltration & backend access
MITM data interception
Inconsistencies & privilege escalations

But, not OpenAI, correct.

1 Like