Who builds QuillBotAI Pro
A free AI text detector is only worth trusting if you know who made it, how it works, and where it fails. This page answers all three.
Who writes and builds this
Zoe Parker
Founder & Lead AI Research Scientist, QuillBotAI Pro
Zoe founded QuillBotAI Pro and leads its detection research. Her work focuses on computational linguistics and on identifying the statistical patterns that large language models leave in text. She writes every article on this site.
NLP Specialization — DeepLearning.AI via Coursera (2024–2025)
LinkedIn profile →How the detector actually works
QuillBotAI Pro is a statistical detector, not a plagiarism checker. It does not compare your text against a database of existing documents. Instead it measures properties that differ, on average, between human and model-generated writing:
- Word predictability (perplexity) — language models tend to choose high-probability next words; human writing wanders more.
- Sentence-length variation (burstiness) — human paragraphs mix long and short sentences; model output is often more uniform.
- Vocabulary diversity — how much the vocabulary range varies across the passage.
- Model-specific phrasing patterns — recurring constructions associated with particular model families.
Results are returned as a per-sentence heatmap rather than a single pass/fail number, so you can see which specific lines drove the score.
What it cannot tell you
No AI detector is 100% accurate, including this one. We publish the limits rather than hide them:
- Short passages (under roughly 50 words) do not carry enough signal to score reliably.
- Heavily edited, paraphrased, or humanized text lowers confidence, sometimes to the point of being unusable.
- Non-native English writing can trigger false positives on any statistical detector. Stanford researchers (Liang et al., 2023) found detectors misclassified a majority of real TOEFL essays as AI-generated.
- A detector result is evidence to weigh, never proof of authorship. It should not be the sole basis for an academic or employment decision.
Why this exists
Most AI detectors either paywall the useful output, cap word counts, or return a single confident-looking percentage with no explanation. That last habit causes real harm when a number gets used to accuse a student. QuillBotAI Pro is free, has no word limit, requires no account, and shows its reasoning per sentence so the result can be argued with.
Read more about the false-positive problem in our write-up on ESL students and detector accuracy, or see the academic integrity guidance for educators.
Privacy and contact
Text and media submitted to the detector are processed in volatile memory and are not intentionally retained. Full details are in the privacy policy.
Corrections, questions, or a detection result you think is wrong — reach out via LinkedIn. We correct published errors rather than quietly editing them away.