Disclosing generative AI use for writing assistance should be voluntary

Area of Science:

• Scientific communication
• Artificial intelligence in research

Background:

• Generative artificial intelligence (GenAI) is increasingly used for manuscript writing support.
• Initial justifications for mandatory GenAI disclosure are becoming less relevant.
• Previous stance favored mandatory disclosure, but this has evolved.

Purpose of the Study:

• To re-evaluate the necessity of mandatory disclosure for GenAI writing assistance.
• To propose an alternative policy for acknowledging GenAI use in scientific manuscripts.

Main Methods:

• Analysis of evolving GenAI capabilities and research practices.
• Assessment of the practicalities and implications of mandatory vs. voluntary disclosure.
• Consideration of ethical concerns, including credit attribution and bias.

Main Results:

• Mandatory disclosure is impractical due to the difficulty in distinguishing human/AI-generated text.
• Concerns exist that mandatory disclosure could introduce bias against non-native English speakers and affect peer review.
• The contribution of GenAI to writing is diminishing to a point not requiring mandatory recognition.

Conclusions:

• A voluntary disclosure policy for GenAI in manuscript writing is now supported.
• Authors should have the discretion to disclose GenAI use, especially where writing is key or beneficial.
• A post-publication voluntary checkbox is suggested as a preferable disclosure mechanism over free-text notes.