Human First Needs an Audit Trail, Not a Checkbox
Every distributor is about to ask you a version of the same question: was this made by a human? A Hypebot op-ed published this week argues that the way the industry is currently planning to ask it produces a claim nobody can check, and that the fix is structural rather than cosmetic.
The Argument in Short
Jeffrey Anthony, director of the Musical Form Institute, is responding to an earlier piece by Cynthia Zhu, who proposed adding a self-attestation checkbox to the upload forms at DistroKid, CD Baby, TuneCore and similar services. Zhu's case was that the checkbox becomes part of the artist-distributor contract, giving it legal weight at close to zero marginal cost.
Anthony's objection is about what happens the moment that claim is challenged.
A distributor that decides an artist lied has to investigate, demand evidence, publish a standard, run an appeals process and potentially litigate. That is not zero marginal cost. And with no industry-wide consensus or federal regulation, a rejected artist simply moves to another distributor, which nullifies the enforcement entirely.
Zhu's own framing supports the point: "The claim is what you see; the enforcement is what makes the claim worth seeing."
What "Human First" Even Means
Before enforcement there is a definitional problem, and it is not trivial. Anthony lists the questions nobody has answered:
- Was the recording made wholly by a human, partly, or through some mix of human and machine processes?
- Do mastering plugins that have been on the market for a decade and are now marketed as AI count as AI?
- What about tools that start from a human performance and then replace the musician's timing and pitch with machine-determined values?
- Is a voluntary sticker from the RIAA, IFPI and A2IM actually assurance of anything?
Human first does not have to stay an undefined marketing claim. But it has to be defined specifically enough to be tested.
The Proposed Alternative
The Musical Form Institute's answer is the Certificate of Embodied Production, or CEP. It is not a voluntary label. It is a certification mark, pending federal registration with the USPTO under Serial No. 99559776, certifying that recordings were produced without grid correction, pitch correction or generative AI.
The model is deliberately borrowed from Certified Organic, Fair Trade and LEED. A certifying organization publishes the standard, evaluates applicants, controls who may use the mark, and maintains procedures for enforcement, suspension and revocation.
The line CEP draws is where authority over the final form of the recording moves away from the musician. Grid correction replaces embodied timing with machine-defined timing. Pitch correction replaces expressed pitch with machine-defined perfect pitch. Generative AI produces material with no embodied musical act at all.
Notably, CEP is not an analog purity argument. Assistive processes including mastering and EQ plugins are allowed, because tools alone do not tell you where authority over musical form sits. Dance music built on a grid is treated differently too, on the reasoning that a dance track finds its completion when a DJ performs it in a particular room at a particular tempo.
Behind the mark sits a published standard, an evidentiary review process, conflict-of-interest controls, a public registry, a permanent audit trail and formal appeals procedures. Every recording is reviewed by a trained human analyst before certification.
Why Independent Artists Should Care
This is a philosophical debate with a very practical consequence.
Verification is becoming a marketing layer. Luminate is rolling out AI Generated, AI and Human tags. Apple Music is labeling AI tracks. Spotify is badging AI Personas from mid-September. ARIA has barred wholly AI-generated recordings from its charts. Deezer reported taking in close to 90,000 fully AI-generated tracks a day in June.
In that environment, being verifiably human stops being a given and becomes a claim you have to be able to support.
The self-attestation model puts the burden on you anyway. If a distributor asks you to tick a box, you are the one carrying the legal exposure for that statement. Anthony's critique is that you carry the risk while the enforcement machinery that would make the claim meaningful does not exist.
Documentation is the hedge. Whatever standard wins, the artists who come through it cleanly will be the ones who kept records.
What to Do Now
- Keep session files and dated stems. Not for aesthetics. For evidence.
- Film process. Writing sessions, tracking, rough takes. It doubles as your best content and your best proof.
- Fill in complete credits. Performer, writer, producer, engineer, session musicians, on every release, everywhere.
- Answer distributor AI disclosures accurately. Partial AI use is not disqualifying under most emerging frameworks. Undisclosed use is what creates exposure.
- Know what your tools do. If your mastering chain contains processes now marketed as AI, that is worth knowing before someone asks.
- Do not lean on absence. No label does not mean verified human. Under Luminate's framework, an absent tag confirms nothing.
The Uncomfortable Middle
Anthony's position raises a question most artists have not sat with. Grid correction and pitch correction are near-universal in commercial production. A standard that excludes them excludes an enormous share of contemporary records, including plenty made entirely by humans in rooms.
That is a real tension, and it is not obvious that the industry will settle on his line rather than a looser one. Others would argue that a performer who chooses to use pitch correction is still exercising authority over the final form, and that the meaningful distinction is generative AI rather than editing.
What is not in dispute is the underlying mechanism. Zhu and Anthony agree that a claim without machinery behind it is marketing language. They disagree about who should own the machinery. Anthony's case that it should sit outside the firms that earn revenue on every upload is the harder one to argue against.
FAQ
Will I be required to declare AI use? Increasingly yes. Major distributors and DSPs are rolling out disclosure fields, and chart bodies including IFPI and ARIA have adopted eligibility principles that depend on them.
Does using autotune make my music AI? Under most current industry frameworks, no. Those target generative AI. The CEP standard draws a stricter line that also excludes pitch and grid correction, but it is one proposed standard, not a rule.
Is a distributor checkbox legally meaningful? It becomes part of your contract with the distributor, so it carries contractual weight. The critique is that contractual weight without independent review, appeals and cross-industry consensus does not produce real enforcement.
Should I pursue certification? It depends on whether provenance is central to your positioning. For acoustic, jazz, folk and live-tracked artists it may become a genuine differentiator. For most producers working in modern pop or electronic formats, accurate credits and honest disclosure will matter more.
What is the low-effort version of protecting myself? Complete credits, dated session files, honest disclosure fields and some process footage. That covers most scenarios regardless of which standard prevails.
Turn Provenance Into Positioning
Documented human work is becoming a marketing advantage, not just paperwork. StreamLord Music Marketing builds that into artist positioning and full campaigns across artist branding, organic Spotify playlisting with no bots, influencer marketing, TikTok strategy and YouTube music video promotion. Start at streamlordmusic.com.
Adapted and expanded from "'Human-First' Needs More Than Just a Checkbox" by Jeffrey Anthony, published on Hypebot, August 27, 2026.