The Grammys ruled an AI voice cannot itself win an award

The Recording Academy's own interview and current rulebook require meaningful human authorship before AI-assisted work can win.

Historical event

June 17, 2023

First source published

June 17, 2023

Site publication

September 18, 2026

What happened

On 17 June 2023, Recording Academy CEO Harvey Mason Jr. discussed new rules for the 2024 Grammy Awards, the 66th ceremony, in an interview published on the Academy's own site. Asked why the process needed to account for artificial intelligence, he said "AI is going to absolutely, unequivocally have a hand in shaping the future of our industry," and described a rule the Academy had just adopted: "Only human creators are eligible to be submitted for consideration for, nominated for, or win a GRAMMY Award." The Academy's current Rules and Guidelines book, as retrieved on 16 September 2026, still carries the same clause under the heading "Generative Artificial Intelligence".

What the documents say

Mason gave a concrete example of how the rule works: "If there's an AI voice singing the song or AI instrumentation, we'll consider it [for submission]. But in a songwriting-based category, it has to have been written mostly by a human... only a human performer can be considered for a GRAMMY." The current rulebook states this more precisely: a work with "no human authorship is not eligible in any categories", while a work containing AI material is eligible provided the human authorship "must be meaningful and more than de minimis" and relevant to the category entered; critically, "the author(s) of any A.I. material incorporated into the work are not eligible to be nominees or Grammy recipients insofar as their contribution... consists of such A.I. material." The rulebook defines de minimis as "lacking significance or importance; so minor as to merit disregard."

Why it matters for makers

The mechanism worth naming is that eligibility is assessed per contribution, not per track. A single recording can contain both an eligible human songwriting credit and an ineligible AI vocal performance at once, under the rulebook's own logic; the two are judged separately rather than the whole work being accepted or rejected as one unit. For a producer building a track around a synthetic voice or AI instrumentation, this means the human-written elements, lyrics, top-line melody, arrangement, may remain eligible even when a model-generated layer does not.

What to check before you use it

Before submitting any work involving AI-generated elements for awards consideration, check which specific categories require which specific kind of human contribution, since the rulebook ties eligibility to the category entered, not to the recording as a whole. Document, contemporaneously, which parts of a composition or performance were human-authored and which were AI-generated, since the "meaningful and more than de minimis" threshold is a judgement call the Academy, not the submitter, ultimately makes. This is an editorial reading of the rulebook's own criteria, not legal advice.

• Which category is being entered, and does its own definition require human songwriting, human performance, or both?
• Can the human-authored portion of the work be documented as more than "de minimis" on its own?
• Does any AI-generated element in the work disqualify only that element, or the submission as a whole?

Read together, the 2023 interview and the current rulebook describe a considered, contribution-by-contribution standard rather than a blanket ban: AI may take part in a Grammy-submitted work, but the Academy's own rule reserves the award itself for whoever did the human part of the job.

Sources & reading trail

Harvey Mason Jr.'s own explanation of why and how the Academy adopted its human-authorship rule for AI-assisted work.

Source published: 17 June 2023 · Retrieved: 16 September 2026

Current living rulebook text of the Generative Artificial Intelligence eligibility clause, quoted verbatim.

Source published: Not established · Retrieved: 16 September 2026

Papers, terms and official documents establish the record; the maker reading and the checks are Signal to Song editorial analysis. This retrospective draft does not imply the site published on the event date.

Continue reading

• A federal bill would give Americans a property right in their voice
• The Copyright Office ruled prompts alone do not make an author
• Open models show two ways to disclose training data
• Browse the complete the archive

Published September 18, 2026, not on the date of the event described.