AI Writing Check

AI Writing Check is a specialized detection utility developed as a joint initiative between Quill.org and CommonLit.org. Built on a technical framework designed to assist educators in the wake of the LLM explosion, the tool leverages the RoBERTa model specifically fine-tuned for identifying text generated by GPT-3 and GPT-3.5 architectures. In the 2026 landscape, while many commercial detectors have shifted toward high-cost subscription models, AI Writing Check maintains its position as a critical open-access resource for the public sector. Its architecture prioritizes a 'human-in-the-loop' philosophy, emphasizing that its probability scores should be used as a conversation starter rather than a definitive verdict. The system analyzes linguistic patterns, perplexity, and burstiness to generate a confidence percentage. Despite the evolution of LLMs toward more human-like variance, this tool remains a benchmark for baseline AI detection in K-12 and higher education, providing a transparent, privacy-conscious alternative to black-box commercial competitors. It remains committed to data privacy, ensuring that student submissions are processed in-memory and not used to train future iterations of commercial models.