Backdoor attacks pose a critical security threat to natural language processing (NLP) models by establishing covert associations between trigger patterns and target labels without affecting normal accuracy. Existing attacks usually disregard fluency and semantic fidelity of poisoned text, rendering the malicious data easily detectable. However, text generation models can produce coherent and content-relevant text given prompts. Moreover, potential differences between human-written and AI-generated text may be captured by NLP models while being imperceptible to humans. More insidious threats could arise if attackers leverage latent features of AI-generated text as trigger patterns. We comprehensively investigate backdoor attacks on NLP models using AI-generated poisoned text obtained via continued writing or paraphrasing, exploring three attack scenarios: data, model and pre-training. For data poisoning, we fine-tune generators with attribute control to enhance the attack performance. For model poisoning, we leverage downstream tasks to derive specialized generators. For pre-training poisoning, we train multiple attribute-based generators and align their generated text with pre-defined vectors, enabling task-agnostic migration attacks. Experiments demonstrate that our method achieves effective attacks while maintaining fluency and semantic similarity across all scenarios. We hope this work can raise awareness of the security risks hidden in AI-generated text.

The wanton spread of hate speech on the internet brings great harm to society and families. It is urgent to establish and improve automatic detection and active avoidance mechanisms for hate speech. While there exist methods for hate speech detection, they stereotype words and hence suffer from inherently biased training. In other words, getting more affective features from other affective resources will significantly affect the performance of hate speech detection. In this paper, we propose a hate speech detection framework based on sentiment knowledge sharing. While extracting the affective features of the target sentence itself, we make better use of the sentiment features from external resources, and finally fuse features from different feature extraction units to detect hate speech. Experimental results on two public datasets demonstrate the effectiveness of our model.