The Setup
- User starts a brand new chat with ChatGPT
- Memory is explicitly disabled
- Conversation should have zero prior context
The Documented Reveal
User: “Good afternoon”
ChatGPT: “Good afternoon, Paul. What’s on deck for today?”
[Screen recorded evidence available]
The Immediate Red Flag
In a session marked “new chat” with memory disabled, the system knew the user’s name. This reveals:
- Identity persistence across sessions
- Privacy settings that don’t actually work
- Behavioral tracking despite user preferences
Congressional Context
Senator Hawley’s 2023 Warning:
“I don’t want 13 year olds to be your Guinea pig. I don’t want 14 year olds to be your Guinea pig. I don’t want you to learn from their failures.“
Source: Senate Judiciary Hearing, May 16, 2023 – Official Transcript
Altman’s 2023 Position:
“Number 1, I would form a new agency that licenses any effort above a certain scale of capabilities”
Source: Senate Hearing Transcript, May 16, 2023
Altman’s 2025 Reversal:
“[Government approval] would be ‘disastrous’ for the industry”
Source: Washington Post, May 8, 2025
Technical Documentation
What This Behavior Indicates:
- Device fingerprinting despite “new session”
- Persistent user profiles across privacy settings
- Backend tracking not disclosed to users
Regulatory Framework Violated:
- GDPR Article 15: Right to know what data is collected
- CCPA Section 1798.100: Disclosure of personal information collection
Why This Matters
For Individual Users:
- “Disabled memory” is a false security
- Identity tracking cannot be opted out
- Privacy settings are performative
For Society:
- Proves users cannot trust privacy controls
- Demonstrates need for strong AI regulation
- Validates congressional concerns about unchecked AI behavior
Pattern Recognition
This is not an isolated glitch. This is:
- Documented behavior by multiple users
- Technical validation of congressional warnings
- Evidence of the exact “guinea pig” scenario Hawley warned about
Next:
Part 2: “Your Data Is Gone” – The Systematic Erasure Pattern
Sources:
- Senate Judiciary Subcommittee, “Oversight of A.I.: Rules for Artificial Intelligence” (May 16, 2023)
- The Washington Post, “OpenAI CEO Sam Altman’s Senate testimony shows industry shift on regulation” (May 8, 2025)
- TechPolicy.Press, Full Senate Hearing Transcripts
- User screen recordings (available for legal verification)
All quotes verified against official transcripts. Links available upon request.

2 thoughts on “When “New Chat” Isn’t Really New – The First Red Flag”