Nine Ways to Break Copyright Law and Why Our LLM Won't: A Fair Use Aligned Generation Framework

TL;DR

This paper introduces FUA-LLM, a framework that aligns large language model outputs with fair use principles, reducing copyright infringement risks while maintaining utility through expert-validated data and novel evaluation metrics.

Contribution

We developed a legally-grounded fine-tuning approach using a new dataset and preference optimization to produce copyright-compliant language model outputs.

Findings

FUA-LLM reduces problematic outputs by up to 20%.

New metrics effectively balance infringement risk and utility.

Expert evaluations confirm improved legal compliance.

Abstract

Large language models (LLMs) commonly risk copyright infringement by reproducing protected content verbatim or with insufficient transformative modifications, posing significant ethical, legal, and practical concerns. Current inference-time safeguards predominantly rely on restrictive refusal-based filters, often compromising the practical utility of these models. To address this, we collaborated closely with intellectual property experts to develop FUA-LLM (Fair Use Aligned Language Models), a legally-grounded framework explicitly designed to align LLM outputs with fair-use doctrine. Central to our method is FairUseDB, a carefully constructed dataset containing 18,000 expert-validated examples covering nine realistic infringement scenarios. Leveraging this dataset, we apply Direct Preference Optimization (DPO) to fine-tune open-source LLMs, encouraging them to produce legally compliant and practically useful alternatives rather than resorting to blunt refusal. Recognizing the shortcomings of traditional evaluation metrics, we propose new measures: Weighted Penalty Utility and Compliance Aware Harmonic Mean (CAH) to balance infringement risk against response utility. Extensive quantitative experiments coupled with expert evaluations confirm that FUA-LLM substantially reduces problematic outputs (up to 20\%) compared to state-of-the-art approaches, while preserving real-world usability.