Detecting AI generated text using neural networks

For humans, distinguishing machine generated text from human written text is men- tally taxing and slow. NLP models have been created to do this more effectively and faster. But, what if some adversarial changes have been added to the machine generated text? This thesis discusses this issue and text detectors in general. The primary goal of this thesis is to describe the current state of text detectors in research and to discuss a key adversarial issue in modern NLP transformers. To describe the current state of text detectors a Systematic Literature Review was done on 50 relevant papers to machine-centric detection in chapter 2. As for the key ad- versarial issue, chapter 3 describes an experiment where RoBERTa was used to test transformers against simple mutations which cause mislabelling. The state of the literature was written at length in the 2nd chapter, showing how viable text detection as a subject has become. Lastly, RoBERTa was shown to be vulnerable to mutation attacks. The solution was found to be fine-tuning it to some heuristics, as long as the mutations can be predicted the model can be fine tuned to detect them