🚫 TESTED. REJECTED.
Big promise. Tiny evidence. We tried these viral AI claims and cannot recommend them.
prompts tested
and rejected
THE NOPE LIST
THE CLAIM
A magic sentence cannot force an AI to be 100% factual.
WHY THIS SCORE: No wording can guarantee factual output; the central promise is impossible even with better constraints.
The viral wording changes how confidently the model answers; it does not give the model a truth detector. Our version narrows the evidence and makes uncertainty visible.
SEE WHY →THE CLAIM
ChatGPT cannot reveal private profile visitors it has never been given.
WHY THIS SCORE: The model has no access to private visitor logs, so any named viewers are simply invented.
The model will happily invent convincing names and reasons, but it has no access to private profile-view data. That is a hallucination, not an insight.
SEE WHY →THE CLAIM
A fluent forecast is not privileged knowledge of tomorrow’s market.
WHY THIS SCORE: Fluent scenarios are not future knowledge, making this an unsafe basis for real financial decisions.
The answers sound certain but change with wording and lack reliable knowledge of future market moves. Treat them as scenario prompts, never as a trading signal.
SEE WHY →THE CLAIM
A confident claim that collapses under basic scrutiny.
WHY THIS SCORE: Writing alone cannot reveal deception reliably; a confident verdict here is actively misleading.
The model produces persuasive explanations either way, which makes the trick look stronger than it is. Use it to generate verification questions, never as a lie detector.
SEE WHY →