Nine Ways to Break Copyright Law and Why Our LLM Won't: A Fair Use Aligned Generation Framework
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866910974941855744 |
|---|---|
| author | Sharma, Aakash Sen Sanyal, Debdeep Srivastava, Priyansh H., Sundar Atreya Karande, Shirish Kankanhalli, Mohan Mandal, Murari |
| author_facet | Sharma, Aakash Sen Sanyal, Debdeep Srivastava, Priyansh H., Sundar Atreya Karande, Shirish Kankanhalli, Mohan Mandal, Murari |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2505_23788 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Nine Ways to Break Copyright Law and Why Our LLM Won't: A Fair Use Aligned Generation Framework Sharma, Aakash Sen Sanyal, Debdeep Srivastava, Priyansh H., Sundar Atreya Karande, Shirish Kankanhalli, Mohan Mandal, Murari Computation and Language Artificial Intelligence 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. |
| title | Nine Ways to Break Copyright Law and Why Our LLM Won't: A Fair Use Aligned Generation Framework |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2505.23788 |