🤖 AI Summary
The proliferation of AI-generated text has blurred the boundary between human- and machine-authored content, posing significant challenges for detection and author attribution—particularly due to strong model dependency and poor scalability of existing methods. To address these issues, we propose WhosAI, the first triplet contrastive learning framework that jointly models detection and author attribution. WhosAI operates in a model-agnostic semantic embedding space, achieved through cross-model alignment and joint representation learning across multiple generative models, enabling end-to-end unified discrimination. Crucially, it supports dynamic incorporation of new AI models without retraining. Evaluated on TuringBench—a large-scale benchmark comprising 200K news articles—WhosAI achieves state-of-the-art performance on both Turing-test-style detection and author attribution tasks, significantly outperforming existing baselines.
📝 Abstract
The significant progress in the development of Large Language Models has contributed to blurring the distinction between human and AI-generated text. The increasing pervasiveness of AI-generated text and the difficulty in detecting it poses new challenges for our society. In this paper, we tackle the problem of detecting and attributing AI-generated text by proposing WhosAI, a triplet-network contrastive learning framework designed to predict whether a given input text has been generated by humans or AI and to unveil the authorship of the text. Unlike most existing approaches, our proposed framework is conceived to learn semantic similarity representations from multiple generators at once, thus equally handling both detection and attribution tasks. Furthermore, WhosAI is model-agnostic and scalable to the release of new AI text-generation models by incorporating their generated instances into the embedding space learned by our framework. Experimental results on the TuringBench benchmark of 200K news articles show that our proposed framework achieves outstanding results in both the Turing Test and Authorship Attribution tasks, outperforming all the methods listed in the TuringBench benchmark leaderboards.