Model-independent measures for authorship discrimination using a Bayesian probabilistic approach

The legal and scientific world is increasingly concerned about their inability to determine and ascertain the identity of the writer of a text. More and more often the question arises as to whether an article or work handed in by e.g. a researcher was actually produced by the alleged author of the questioned text. The role of artificial intelligence (AI) is increasingly debated due to its dangers of undeclared use. A current example is undoubtedly the undeclared use of ChatGPT to write texts. The ISA Lecture promotes an AI model-independent measure to support discrimination between hypotheses on authorship of various multilingual texts written by humans or produced by intelligence media. The syntax of texts written by humans tends to differ from that of texts produced by AIs. This difference can be grasped and quantified even with short texts. To meet the efficiency criteria required for the evaluation of forensic evidence, a probabilistic approach is implemented. To offer a consistent classification criterion, a metric called Bayes factor is implemented. The proposed probabilistic method represents an original approach in stylometry. Analyses performed over multilingual texts covering different scientific and human areas of interest reveal the feasibility of a successful authorship discrimination with limited misclassification rates. Controversy of literature originality can also be approached in total respect of standards for evaluative reporting and legal jurisprudence.

Speaker

ISA Visiting Fellow - Franco Taroni

Full Professor Université de Lausanne, École des sciences criminelles

Visit Prof. Taroni's website