We welcome you to CDS-KIAC talk on 15th December 2025 (Monday). The details are as below:
Speaker: Dr. Tuhin Chakrabarty, Stony Brook University, US
Title: Readers Prefer Outputs of AI Trained on Copyrighted Books over Expert Human Writers
Date and Time: December 15, 2025: 11:00 AM
Venue: The Seminar will be held on HYBRID Mode
# 102 CDS Seminar Hall /MICROSOFT TEAMS
Please click on the following link to join the Seminar:
Meeting link
Abstract:
The use of copyrighted books for training AI models has led to numerous lawsuits from authors concerned about AI’s ability to generate derivative content. Yet it’s unclear whether these models can generate high-quality literary text while emulating authors’ styles and voices. To answer this, we conducted a preregistered study comparing MFA-trained expert writers with three frontier AI models (ChatGPT, Claude, and Gemini) in writing up to 450-word excerpts emulating the diverse styles of 50 award-winning authors. Based on blind pairwise evaluations by 159 representative expert and lay readers, we found that when AI was simply prompted to imitate an author, experts overwhelmingly preferred human writing—both for capturing the author’s style and for overall quality. But when ChatGPT is finetuned on an author’s complete works, these results flipped dramatically: experts now preferred the AI-generated text on both measures. Lay readers showed similar shifts. Fine-tuning eliminates detectable AI stylistic quirks like overuse of clichés that readers find off-putting in standard AI output, which largely explains this reversal. These fine-tuned outputs also fooled AI detectors 97% of the time, in addition to being 99.7% cheaper compared to typical professional writer compensation. Together, these results have direct implications for ongoing copyright lawsuits, specifically the legal question of whether AI training harms the market for authors’ work.
Bio of Speaker:
Tuhin Chakrabarty is Assistant Professor in the Computer Science Department at Stony Brook University (SUNY). Prior to this, he obtained his PhD from Columbia University. His research interests are broadly in AI, NLP, and Human-AI Interaction, and his work often relies on knowledge, methods, and perspectives from multiple disciplines to address complex problems that cannot be fully understood or solved within the boundaries of Computer Science. Tuhin’s work has been covered in MIT Tech Review, Bloomberg, and the Washington Post, and he has received a Best Paper Honorable Mention award at ACM CHI and an Outstanding Position Paper award at ICML. Most recently, his research on Generative AI and Fair Use is being profiled by The New Yorker, and his empirical findings on market dilution and substitution effects are being used by leading U.S. law firms in ongoing AI copyright litigation.
About: This talk is part of the KIAC Seminar Series on AI for Fintech.
Host Faculty: Danish Pruthi, CDS
ALL ARE WELCOME
