Show, Don't TELL: Explainable AI-Generated Text Detection

2026-05-26

AI Summary

This work addresses the limitation of existing AI-generated text detectors, which provide uninterpretable scores without actionable justifications, failing to meet users’ demands for transparent and verifiable evidence. To bridge this gap, the authors propose TELL—an architecture that reframes detection as a human-centered explainability problem. TELL natively outputs both a detection score and fine-grained textual rationales highlighting specific spans deemed indicative of AI or human authorship, thereby “showing rather than telling” to support users’ critical evaluation. The model is trained on domain-specific authorship-labeled data and optimized via curriculum learning combined with GRPO reinforcement learning. Experimental results demonstrate that TELL achieves state-of-the-art AUROC performance (0.927) while generating explanations that outperform baselines in specificity and falsifiability, attaining a 72.3% win rate in human evaluations.

Abstract

Research on AI-generated text detection has presented a number of approaches to discern human from AI prose, some of which achieving high in-distribution performance. However, real-world applicability has stalled because their outputs are misaligned with the needs of users, such as professors, who are presented with a numeric score that has no attached explanation. We tackle this issue with a novel architecture, TELL, that bakes explainability from the ground-up. While our system still offers a numerical score like other detectors for comparability, TELL takes a fundamentally different approach where we aim to show the user the "tells" by which the model believes a text is AI or human-written, to empower the user to decide who wrote a text using their own judgment and understanding of the context of the writing and its alleged author. We train TELL on a custom SFT dataset of domain-specific authorship annotations, and further refine the system using GRPO with curriculum learning to improve performance. We achieve competitive performance with state-of-the-art detectors (AUROC 0.927) while natively providing annotations that explain the basis for the detector's decision. We further evaluate the quality of our explanations using a dataset of human annotations and report a high (mean 72.3%) win-rate on annotation concreteness, falsifiability, coherence, plausibility and grounding, allowing users to critically think and decide for themselves. Our work thus reframes the problem of AI-generated text detection in a human-centric perspective and paves the way for a new family of detectors that focus on native explainability.

Problem

• AI-generated text detection
• explainability
• human-centric AI
• interpretability
• trustworthy AI

Innovation

• Explainable AI
• AI-generated text detection
• Native explainability
• Human-centric AI
• GRPO