Alignment Whack-a-Mole : Finetuning Activates Verbatim Recall of Copyrighted Books in Large Language Models
Frontier LLM companies have repeatedly assured courts and regulators that their models do not store copies of training data. They further rely on safety alignment strategies via RLHF, system prompts, and output filters to block verbatim regurgitation of copyrighted works, and have cited the efficacy of these measures in their legal defenses against copyright infringement claims. We show that finetuning bypasses these protections: by training models to expand plot summaries into full text, a task
Record details
Published: 21 March 2026
Source: arXiv
Category: Research
Topics: Regulation · Safety & alignment · Copyright & IP
Retrieved: 14 July 2026
Related evidence
These records share source-supplied organisations, an exact publisher byline, automatic topics or regions. The reason is shown on every link; related does not mean supporting, agreeing with or verifying this record.
From Licensing to Open Access: Designing a Sustainable Transition in Operational Weather Data
arXiv · 20 May 2026
SafeMed-R1: Clinician-Audited Safety and Ethics Alignment for Medical Large Language Models
arXiv · 27 May 2026
Challenges to Grassroots Organization Engagement with AI Policy
arXiv · 18 June 2026
Before the Tool Call: Deterministic Pre-Action Authorization for Autonomous AI Agents
arXiv · 21 March 2026
Too much of a good thing? Entrepreneurial orientation and the non-linear governance effects of SaaS platforms
arXiv · 22 March 2026
Why AI Readiness Is an Organizational Learning Problem, Not a Technology Purchase
arXiv · 22 March 2026
How to cite this record
ethics.ai (21 March 2026), “Alignment Whack-a-Mole : Finetuning Activates Verbatim Recall of Copyrighted Books in Large Language Models,” evidence record 6911, https://ethics.ai/record/6911 (originally published by arXiv).
Use and limitations
This page is a stable index and citation surface for a source record. ethics.ai did not author the underlying report and has not independently verified every claim. Automatic topics may be imperfect. For consequential use, quote and cite the original publisher.