Straight answers on AI music and detection
Articles on how AI music detection actually works, how the major generators differ, what streaming platforms really enforce, and where copyright law currently stands. No inflated accuracy claims, and no pretending the hard cases are simple.
Looking for tools rather than reading? See how this detector compares with Ircam Amplify, Deezer and AI Voice Detector — what each one detects, what it costs, how it handles your data, and whether its accuracy claim is actually backed by a published method.
Asking about a voice rather than a song? How to tell if a voice is AI generated — which cues actually hold up, and why a music-tuned engine like this one isn't the right tool for a bare voice clip.
Guide
Guide
How to Check a Song for AI: An Evidence-Led Workflow
A step-by-step method combining ears, measurements and provenance — in the order that actually works.
9 min read
Guide
Reading Audio Metadata Without Overclaiming What It Proves
Encoder strings and ISRCs are the only origin clues a file carries — and most files carry none at all.
12 min read
Guide
Spectrogram Myths: What the Image Can Actually Establish
The high-frequency cutoff everyone points to as AI proof is usually just an MP3 doing its job.
12 min read
Guide
How to Triage a Music Catalogue Without Automating Accusations
At realistic incidence rates most flagged tracks turn out human. Here's how to run a batch screen properly anyway.
12 min read
Guide
Loudness Is Not Provenance: Reading LUFS in Context
"Over-compressed, so it's AI" is one of the weakest claims in the field. Here's what a loudness meter actually shows.
12 min read
Guide
Stereo Width Is a Mix Diagnostic, Not an AI Test
Stereo width is the easiest thing in a mix to fake, which makes it the shakiest thing to build a provenance claim on.
12 min read
Guide
What Repetition Reveals—and What It Cannot
Copy-and-paste is a production technique, not a confession. Here's what a repetition reading actually shows.
12 min read
Guide
Grid-Tight Timing Does Not Prove AI Authorship
Quantisation has been ordinary studio craft for forty years. Here's what grid-perfect timing does and doesn't prove.
12 min read
Guide
Perfect Pitch Is Not an AI Fingerprint
Pitch correction has been routine since 1997. Here's what grid-exact intonation actually shows, and what it can settle.
12 min read
Comparison
Comparison
Suno or Udio? Why the Production Workflow Matters More
Same goal, different machinery — and different tells left behind in the audio.
8 min read
Comparison
Detector or Trained Ear? Give Each One the Right Job
Two very different instruments, and two very different ways of getting it wrong.
11 min read
Comparison
Synthetic and Recorded Vocals: A Careful Listening Framework
The vocal tells that still hold up in 2026 — and the ones that quietly stopped working.
11 min read
Legal
Legal
AI Music Copyright: A Map of Clear and Unsettled Questions
Authorship, training data and voice likeness, sorted into settled and unresolved.
9 min read
Legal
Monetising AI Music: Rights, Platforms and Practical Risk
Streaming, sync and sales, mapped against where the payout actually stops.
11 min read
Technical
Explainer
Industry
Industry
Does Spotify Detect AI Music? Separate Policy from Audio Analysis
Moderation, fraud checks and disclosure rules — a different job from acoustic detection.
8 min read
Industry
YouTube and AI Music: Which Rule Applies to Whom?
Disclosure duties, auto-labels and the inauthentic-content rule, drawn straight from YouTube's own policy pages.
12 min read
Industry
A Four-Stage AI Music Intake Process for Rights Teams
Listening alone stopped working on polished hybrids. Here's the paperwork and triage that still does the job.
12 min read