← The AI Hype Audit — all 449 verdicts

BS

Straight BS: the 'professor showed us how to bypass AI detectors' reel is a humanizer promo where the only '100% Human' score comes from Verva's own built-in checker, not Turnitin or GPTZero

Facebook · Oct 2026

The claimOn-screen: 'POV: Professor randomly just showed us how to bypass AI detectors.' A man at a lecture-hall podium types verva.com (tab title 'Verva, formerly TwainGPT'), pastes a short ChatGPT paragraph about caffeine into the Humanizer, and the result shows a green '100% Human' badge. Caption: 'random tuesday lecture'.

Verva is real. It is the rebranded TwainGPT, an AI humanizer and detector sold by Overthink, LLC, with plans listed on its site at $8, $20 and $40 a month, and it markets itself as 'built for GPTZero, ZeroGPT, Copyleaks, Quillbot, Turnitin.' What the reel does not show is any of those detectors. The '100% Human' badge at the end is Verva's own score on Verva's own output, the same company grading its own homework. Nobody pastes the text into Turnitin or GPTZero on camera. The 'professor' framing is a skit: the presenter is logged into a paid Verva dashboard ('Welcome, David', about 28,000 words remaining) and types the brand's URL slowly enough for the camera to read it, which is how sponsored humanizer reels are built. Independent results on humanizers are mixed at best: some passes slip past some detectors some of the time, and the detectors keep updating. Turnitin shipped dedicated 'AI bypasser' detection in August 2025 aimed at exactly this kind of rewritten text. And the bigger risk is not the score. Submitting AI-written work run through a humanizer is still an academic-integrity violation at most schools, whether or not a detector flags it, and a humanized essay you cannot explain in office hours is worse evidence than a flagged one.

What holds up

  • Reel read frame by frame: the address bar shows verva.com, the tab reads 'Verva, formerly TwainGPT', the dashboard greets 'David' with roughly 28,024 humanizer words left, and the final frame is Verva's own '100% Human' badge under an 84-word paragraph about caffeine.
  • verva.com read directly: AI humanizer plus AI detector, 'formerly TwainGPT', claims to be 'built for GPTZero, ZeroGPT, Copyleaks, Quillbot, Turnitin, Writer, Grammarly', plans listed at $8, $20 and $40 a month; third-party listings name Overthink, LLC as the maker.
  • Turnitin announced AI bypasser detection in late August 2025 (Plagiarism Today, Campus Technology) to flag text modified by humanizer tools; GPTZero publishes its own work on detecting humanized text. No detector outside Verva appears in the reel.
  • x-ref #271: same 'pass all AI detectors' promise, same pattern of the tool grading itself.

What doesn’t

  • The only proof is the vendor's own detector scoring the vendor's own output.
  • Staged 'professor' skit with a paid account and a slow-typed brand URL; it reads as an ad, with no paid-partnership label visible in the frames read.
  • Selling detector evasion to students: using it on graded work is an integrity violation regardless of the score.

The catch

The humanizer exists and charges a monthly fee; the reel proves only that Verva likes Verva's rewrite. Turnitin specifically looks for humanizer output now, and a skit in a lecture hall is not a professor's endorsement.

How to actually do it

  • If your course allows AI help, follow its disclosure rule and say which tool you used and how; that is the only route that holds up if a paper is questioned.
  • If you want writing that reads as yours, draft it yourself and use AI for outlines, feedback and grammar, then keep your version history (Google Docs or Word) as proof of process.
  • If you are falsely flagged, ask for a meeting, bring your drafts and edit history, and cite the detector vendor's own guidance that a score alone is not proof.

Verva (formerly TwainGPT) is a real paid humanizer, but the reel never tests it against Turnitin, GPTZero or any outside detector. The professor setup is a staged promo, and humanized AI work is still misconduct under most school policies.

Confidence
High
Posted by
a small college-humor reel page (this reel ~95k plays); reel sent to Buddy Mon Oct 5, 2026, ~9:41 PM

See the original claim →

We test hype for free. We build the real thing for a living.

Thirty minutes, no pitch — and you'll leave with something useful either way.

Book a call with Todd or start with the free Business Checkup →

The Verdict Weekly

Three verdicts every Friday. Free forever, unsubscribe anytime, no spam — that would be ironic.

© Schreier Group · schreiergroup.com · See a wild AI claim? Drop it here and we'll test it.