Scripli Blog
Writing, integrity, and the future of authentic work
Insights on academic integrity, AI writing tools, human authorship proof, and what institutions, educators, and students need to know right now.
AI Detectors Can't Be Fixed. The Approach Is the Problem.
The false-positive rate isn't a bug that will eventually be patched. Output-based detection is fundamentally incapable of answering whether a human wrote something — and that gap will only widen.
Why the Whole Detection Approach Was Backwards All Along
A research paper on image provenance makes a point that applies directly to writing: the internet has been trying to detect fakes when it should have been verifying origins. That is a very different idea.
Your Work Can Be 100% Original and Still Fail an AI Authorship Check
AI-generated text is original — it matches nothing in a plagiarism database. That means passing a plagiarism check no longer tells you, or anyone else, whether a human wrote the work.
AI Detectors Put Innocent Students at Risk. Here's What Actually Works.
AI detection results are probabilistic, not proof. Institutions that act on them expose themselves to legal and reputational risk. Here are the strategies that actually hold up.
Once You're Accused of Using AI, It's Too Late to Prove You Didn't
When a detector flags your writing, the burden falls on you to explain it. But explaining isn't the same as proving. The only answer that actually works has to come before anyone asks.
Scripli Wrapped: The Year Your Writing Finally Talks Back
You spent a year showing up to the page. By December it's hard to see the shape of it. Did you grow? Were you consistent? Scripli Wrapped answers those questions.