Amend copyright licences to halt AI misuse and reassert human control

Nature正文已收录本站
  • CORRESPONDENCE

By

  • Daniel L. Gardner1 &
  • Katta Spiel2
  1. Katie Seaborn
    1. University of Cambridge, Cambridge, UK.

  2. Daniel L. Gardner
    1. Robert Gordon University, Aberdeen, UK.

  3. Katta Spiel
    1. Vienna University of Technology, Vienna, Austria.

Generative artificial-intelligence models, notably large language models (LLMs), are disrupting intellectual-property rights. Crucially, present-day copyright licences cannot accommodate human adaptation and prevent AI misuse simultaneously. Copyright holders can, for instance, select the BY-ND ‘no derivatives’ Creative Commons (CC) licence to protect against AI misuse. However, this also prevents legitimate creative human adaptation of the copyrighted material.

Nature 656, 1070 (2026)

doi: https://doi.org/10.1038/d41586-026-02633-y

Competing Interests

The authors declare no competing interests.

  • AI firms must play fair when they use academic data in training

  • Intellectual property and data privacy: the hidden risks of AI

  • Has your paper been used to train an AI model? Almost certainly

Subjects

Latest on:

  • Intellectual-property rights
  • Publishing
  • Law
  • Transparent research: can big tech learn from big pharma?

    Editorial

  • What incentives do companies need to publish research?

    Nature Index

  • Here’s why China’s science and innovation model is thriving

    Nature Index

  • The future of peer review requires AI support, not AI bans

    Correspondence

  • Sleuth identifies dozens of studies that used the wrong antibody

    News

  • Staggering 90% of biomedical papers now show signs of AI help

    News

  • South Africa’s scientists must register with official body or risk prison, according to draft law

    News

  • The forgotten 20 million: why is the world neglecting Afghanistan’s girls and women?

    Editorial

  • How I use statistics and my law degree to fight human-rights abuses

    Career Feature