Opinion | An Accountability Test for AI

James Taranto is right that artificial intelligence is a tool, not an authorial verdict (“AI: Ghostwriter in the Machine,” op-ed, Sept. 19). The more useful question isn’t whether AI assisted with a piece, but whether the person whose name appears on it can honestly claim its ideas, verify its facts and defend its conclusions.

Writers have always relied on editors, researchers and colleagues to sharpen their work. AI can serve some of those functions with remarkable speed. But assistance becomes substitution when the named author contributes little more than a prompt and an approval—or presents borrowed expertise, experiences or convictions as his own.

This distinction matters because authorship isn’t merely credit for arranging words. It is an act of accountability. A byline tells readers: I mean this. I have examined it. I will answer for it.

No detector can reliably determine whether that promise has been kept. Editors and readers will still need what Mr. Taranto prescribes: human judgment. But that judgment should be guided by a simple principle—AI may help us find and express our voice; it shouldn’t become a means of impersonating one.

Seth Eisenberg

Fort Lauderdale, Fla.

Mr. Eisenberg is a former chairman of the National Writers Union.