Ever wondered if AI could pinpoint who penned a particular piece of writing? Authorship verification, the task of determining if two texts share the same author, has long been a complex puzzle. Traditional methods, relying on analyzing writing styles, have had their limitations. Now, a groundbreaking approach called InstructAV is changing the game. Imagine an AI detective meticulously examining texts, not just for their content, but for hidden stylistic fingerprints. InstructAV uses large language models (LLMs) like cutting-edge AI assistants, and fine-tunes them with a clever technique called Parameter-Efficient Fine-Tuning (PEFT). This allows the model to become an expert at spotting subtle clues in writing style that reveal authorship. What sets InstructAV apart is its ability to explain its reasoning. Instead of just giving a 'yes' or 'no' answer, it lays out the evidence. It points to specific linguistic features—like sentence structure, word choice, and even punctuation—that point to a common author or reveal distinct writing styles. Researchers tested InstructAV on datasets ranging from lengthy IMDB reviews to short, snappy tweets. The results? Impressive accuracy and surprisingly insightful explanations. InstructAV not only outperformed traditional methods but also some of the latest LLM techniques. This breakthrough opens exciting possibilities. InstructAV could be used in forensics to verify the authenticity of documents, in literature to analyze writing styles and influences, and even to improve the detection of plagiarism. While there are still challenges, such as improving the speed of generating explanations, InstructAV represents a big step forward in making AI's judgments clearer and more reliable in authorship verification.