Direct answer
Direct Answer - Turnitin does not explicitly list Doubao or Qwen among the models its AI writing detection tool can identify. Its published documentation names GPT-3, GPT-3.5, GPT-4, and ChatGPT as detectable models, but states that the list is not exhaustive [1]. Whether Turnitin flags text from Doubao (ByteDance) or Qwen (Alibaba) depends on whether those models' output patterns were represented in the training data used to build the detection model. Because both models are relatively newer and primarily serve Chinese-language users, their inclusion in Turnitin's training corpus is uncertain. The safest assumption for any student using these models is that detection is possible and unpredictable.
Why Does Turnitin Flag Certain AI Models and Not Others?
Turnitin's AI writing detection tool works by analyzing text for statistical patterns and markers that are characteristic of machine-generated content [2]. The detection model is trained on a large corpus that includes both human-written academic text and AI-generated samples from various large language models. When a model's writing style — its typical sentence structures, word choice distributions, and coherence patterns — closely matches what the detector learned during training, that text is more likely to be flagged [2].
Turnitin acknowledges that detection coverage varies across models. Models whose output was heavily represented in the training set, such as GPT-4 and ChatGPT, are detected with higher accuracy. Newer or less widely deployed models may fall outside the detector's known pattern library, leading to lower flag rates [2]. Turnitin continuously updates its detection model to keep pace with the rapid evolution of AI language models, but there is always a lag between a model's public release and its inclusion in detection training data. This means that detection capability is not static — a model that evades detection today may be caught after a future update.
Can Turnitin Detect Chinese-Origin AI Models Like Doubao and Qwen?
Turnitin's AI detection primarily targets English-language academic text but has expanded to support detection across multiple languages, including Chinese [3]. However, the accuracy of detection varies significantly by language and by model. Doubao, developed by ByteDance, and Qwen, created by Alibaba, are Chinese-origin large language models that generate text primarily in Chinese (and to a lesser extent in English). These models operate on different training data, tokenization strategies, and output distributions compared to Western models like GPT-4 or Claude [3].
Because Turnitin does not publicly disclose the complete list of models in its detection training corpus, it is impossible to say with certainty whether Doubao and Qwen are currently detectable. The company's research has shown that AI detection accuracy is generally higher for English text than for non-English text, and that models generating output in languages underrepresented in training data may have lower detection rates [3]. Students using Doubao or Qwen to generate Chinese-language academic text should be aware that detection is possible but not guaranteed, and that Turnitin's detection capabilities continue to evolve as new models emerge.
What Should You Do If Your Doubao or Qwen Content Is Flagged by Turnitin?
If Turnitin flags text you generated using Doubao or Qwen, the first step is to review the AI writing report to identify exactly which sections were flagged and at what confidence level [4]. Turnitin's report highlights specific passages that the detector considers likely AI-generated, allowing you to see where the patterns were most pronounced. Many institutions treat an AI flag as a starting point for discussion rather than a definitive judgment of misconduct [4].
Once you understand what triggered the flag, you have several options. You can rewrite the flagged sections manually to incorporate more of your own voice and sentence variation. You can also discuss the findings with your instructor, as some institutions allow students to explain their writing process or provide drafts as evidence of original work [4]. If you need a more reliable solution, using a dedicated AI humanizer — a tool designed to rewrite AI-generated text so that it reads as fully human-written — can help reduce or eliminate the AI score before you submit. The most effective approach is to address the detection issue proactively rather than submitting flagged content and dealing with the consequences afterward.
If you have already written your assignment using Doubao, Qwen, or any other AI model and are concerned about Turnitin detection, you can humanize your text to bring the AI score down to *% or even 0%. Turnitin0's AI humanizer is designed specifically to bypass Turnitin AI detection while preserving your original meaning, academic quality, and document formatting.
※ Turnitin0.com - AI Humanizer Bypassing Turnitin AI Detector
FAQ
Q: Does Turnitin specifically mention Doubao or Qwen in its documentation?
A: No. Turnitin's published documentation names GPT-3, GPT-3.5, GPT-4, and ChatGPT as detectable models but does not specifically mention Doubao or Qwen [1]. The company states its list of detectable models is not exhaustive, meaning the absence of a model from the documentation does not guarantee it cannot be detected.
Q: Is Turnitin more accurate at detecting AI-generated English text than Chinese text?
A: Yes. Research conducted by Turnitin indicates that AI detection accuracy is generally higher for English-language text than for non-English text [3]. Chinese-origin models like Doubao and Qwen that generate text primarily in Chinese may be detected at lower rates, though this varies by model and by the specific content.
Q: Can Turnitin detect AI content after it has been humanized?
A: A high-quality AI humanizer that rewrites text with varied sentence structure, natural word choices, and human-like coherence patterns can significantly reduce or eliminate Turnitin's AI flag. Turnitin0's AI humanizer is specifically designed to bypass Turnitin AI detection while preserving academic quality and original meaning.
Q: What happens if my instructor sees an AI flag on my submission?
A: Institutional policies vary. Many universities treat an AI flag as a discussion prompt rather than automatic evidence of misconduct [4]. Your instructor may ask you to explain your writing process, provide drafts, or resubmit after revision. Being proactive about addressing detection before submission is always the safer approach.
Q: Does Turnitin detect content from all AI models equally?
A: No. Detection accuracy varies depending on whether the specific model's output patterns were included in Turnitin's training data [2]. Models that are well-represented in the training corpus (like GPT-4 and ChatGPT) are detected more reliably. Newer, less common, or non-English-focused models like Doubao and Qwen may have lower detection rates, though this can change as Turnitin updates its detection model.