Fix your writing
with a single click

The ethics of machine writing

Large language models now draft essays, emails and news articles that is indistinguishable from human prose. As these systems spread through classrooms, their affect on academic integrity has become urgent. Recent surveys suggest nearly one in three students has submitted machine-assisted work without disclosure, and instructors report rising uncertainty about what authorship even means.

Detection alone is not a policy. Universities that pair detection with clear disclosure rules see fewer disputes and better outcomes for students who rely on assistive tools for accessibility. The most effective programs treat integrity as something taught, not merely enforced.

Institutions are responding unevenly. Some ban generative tools outright, while others utilize disclosure statements modeled on funding acknowledgments, and a growing minority now teach prompt literacy alongside citation practice. The result is a patchwork of expectations that shifts from course to course, leaving students to guess where the lines is drawn.

This paper argues that authorship is best understood as a spectrum of contribution rather than a binary, and that workable policy should track disclosure rather than detection alone. Drawing on three semesters of survey data, it proposes a framework that