From Perceptions To Evidence: Detecting AI-Generated Content In Turkish News Media With A Fine-Tuned Bert Classifier
A new AI detector exposes how much news is actually written by machines.
Deep Dive
A new study uses a fine-tuned Turkish BERT model to detect AI-generated content in news media for the first time. The model, trained on 3,600 articles, achieved a 0.9708 F1 score. When deployed on over 3,500 unseen articles from 2023-2026, it found an estimated 2.5% of the examined Turkish news content had been rewritten or revised by LLMs, moving beyond perception to empirical evidence.
Why It Matters
This provides the first hard data on AI's real-world infiltration into non-English newsrooms, a critical trust issue.