Interpretable Text Classification Applied to the Detection of LLM-generated Creative Writing

We consider the problem of distinguishing human-written creative fiction (excerpts from novels) from similar text generated by an LLM. Our results show that, while human observers perform poorly (near chance levels) on this binary classification task, a variety of machine-learning models achieve accuracy in the range 0.93 - 0.98 over a previously unseen test set, even using only short samples and single-token (unigram) features. We therefore employ an inherently interpretable (linear) classifier (with a test accuracy of 0.98), in order to elucidate the underlying reasons for this high accuracy. In our analysis, we identify specific unigram features indicative of LLM-generated text, one of the most important being that the LLM tends to use a larger variety of synonyms, thereby skewing the probability distributions in a manner that is easy to detect for a machine learning classifier, yet very difficult for a human observer. Four additional explanation categories were also identified, namely, temporal drift, Americanisms, foreign language usage, and colloquialisms. As identification of the AI-generated text depends on a constellation of such features, the classification appears robust, and therefore not easy to circumvent by malicious actors intent on misrepresenting AI-generated text as human work.

Using Machine Learning to Distinguish Human-written from Machine-generated Creative Fiction

Following the universal availability of generative AI systems with the release of ChatGPT, automatic detection of deceptive text created by Large Language Models has focused on domains such as academic plagiarism and "fake news". However, generative AI also poses a threat to the livelihood of creative writers, and perhaps to literary culture in general, through reduction in quality of published material. Training a Large Language Model on writers' output to generate "sham books" in a particular style seems to constitute a new form of plagiarism. This problem has been little researched. In this study, we trained Machine Learning classifier models to distinguish short samples of human-written from machine-generated creative fiction, focusing on classic detective novels. Our results show that a Naive Bayes and a Multi-Layer Perceptron classifier achieved a high degree of success (accuracy > 95%), significantly outperforming human judges (accuracy < 55%). This approach worked well with short text samples (around 100 words), which previous research has shown to be difficult to classify. We have deployed an online proof-of-concept classifier tool, AI Detective, as a first step towards developing lightweight and reliable applications for use by editors and publishers, with the aim of protecting the economic and cultural contribution of human authors.

Moving Healthcare AI-Support Systems for Visually Detectable Diseases onto Constrained Devices

Image classification usually requires connectivity and access to the cloud which is often limited in many parts of the world, including hard to reach rural areas. TinyML aims to solve this problem by hosting AI assistants on constrained devices, eliminating connectivity issues by processing data within the device itself, without internet or cloud access. This pilot study explores the use of tinyML to provide healthcare support with low spec devices in low connectivity environments, focusing on diagnosis of skin diseases and the ethical use of AI assistants in a healthcare setting. To investigate this, 10,000 images of skin lesions were used to train a model for classifying visually detectable diseases (VDDs). The model weights were then offloaded to a Raspberry Pi with a webcam attached, to be used for the classification of skin lesions without internet access. It was found that the developed prototype achieved a test accuracy of 78% and a test loss of 1.08.

Investigating Markers and Drivers of Gender Bias in Machine Translations

Implicit gender bias in Large Language Models (LLMs) is a well-documented problem, and implications of gender introduced into automatic translations can perpetuate real-world biases. However, some LLMs use heuristics or post-processing to mask such bias, making investigation difficult. Here, we examine bias in LLMss via back-translation, using the DeepL translation API to investigate the bias evinced when repeatedly translating a set of 56 Software Engineering tasks used in a previous study. Each statement starts with 'she', and is translated first into a 'genderless' intermediate language then back into English; we then examine pronoun-choice in the back-translated texts. We expand prior research in the following ways: (1) by comparing results across five intermediate languages, namely Finnish, Indonesian, Estonian, Turkish and Hungarian; (2) by proposing a novel metric for assessing the variation in gender implied in the repeated translations, avoiding the over-interpretation of individual pronouns, apparent in earlier work; (3) by investigating sentence features that drive bias; (4) and by comparing results from three time-lapsed datasets to establish the reproducibility of the approach. We found that some languages display similar patterns of pronoun use, falling into three loose groups, but that patterns vary between groups; this underlines the need to work with multiple languages. We also identify the main verb appearing in a sentence as a likely significant driver of implied gender in the translations. Moreover, we see a good level of replicability in the results, and establish that our variation metric proves robust despite an obvious change in the behaviour of the DeepL translation API during the course of the study. These results show that the back-translation method can provide further insights into bias in language models.