
  <rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
    <channel>
      <title>We are the music makers</title>
      <link>https://wearethemusicmakers.co/latest</link>
      <description>Music tech and AI in music: the news that matters, what we think about it, and a directory of the tools.</description>
      <language>en-us</language>
      <lastBuildDate>Mon, 26 Oct 2026 03:30:00 GMT</lastBuildDate>
      <atom:link href="https://wearethemusicmakers.co/tags/ai-detection/feed.xml" rel="self" type="application/rss+xml"/>
      
  <item>
    <guid>https://wearethemusicmakers.co/opinion/publish-detector-error-rates-not-thresholds</guid>
    <title>The case for keeping AI detectors secret is strong. It still doesn&#39;t cover error rates</title>
    <link>https://wearethemusicmakers.co/opinion/publish-detector-error-rates-not-thresholds</link>
    <description>The labels&#39; subpoena fight with DistroKid has revived calls for distributors to publish everything about their AI detectors. The best counter-argument, that disclosure helps spammers beat the filter, is right about methods and wrong about error rates. Here is where we think the line goes.</description>
    <pubDate>Mon, 26 Oct 2026 03:30:00 GMT</pubDate>
    <dc:creator>We are the music makers</dc:creator>
    <category>ai-detection</category><category>ai-music</category><category>distribution</category><category>distrokid</category><category>policy</category><category>format-counter-take</category>
  </item>

  <item>
    <guid>https://wearethemusicmakers.co/learn/audio-watermarking-explained</guid>
    <title>Audio watermarking explained: how SynthID and AudioSeal mark AI audio, and how easily the marks break</title>
    <link>https://wearethemusicmakers.co/learn/audio-watermarking-explained</link>
    <description>AI audio watermarks hide a signal in the sound that a detector can find later. Here is how the main systems work, what Google and Meta claim, what independent benchmarks found when they attacked them, and what a positive or negative result can and cannot tell you about a track.</description>
    <pubDate>Sat, 24 Oct 2026 06:30:00 GMT</pubDate>
    <dc:creator>We are the music makers</dc:creator>
    <category>ai-detection</category><category>provenance</category><category>watermarking</category><category>ai-music</category>
  </item>

  <item>
    <guid>https://wearethemusicmakers.co/learn/ai-flag-appeal-kit</guid>
    <title>Your human-made track was flagged as AI: an appeal kit, distributor by distributor</title>
    <link>https://wearethemusicmakers.co/learn/ai-flag-appeal-kit</link>
    <description>None of the big three distributors publishes an appeal process for an AI flag. Here is what DistroKid, TuneCore and CD Baby each say in their own help pages as of 24 Oct 2026, the evidence pack to build before you write, a reply you can adapt, and the moves that make a wrong flag worse.</description>
    <pubDate>Sat, 24 Oct 2026 03:30:00 GMT</pubDate>
    <dc:creator>We are the music makers</dc:creator>
    <category>ai-music</category><category>ai-detection</category><category>distribution</category><category>distrokid</category><category>workflows</category>
  </item>

  <item>
    <guid>https://wearethemusicmakers.co/news/2026/10/how-distributors-detect-ai-false-positives</guid>
    <title>How distributors decide your song is &quot;AI&quot;, and what to do when they get it wrong</title>
    <link>https://wearethemusicmakers.co/news/2026/10/how-distributors-detect-ai-false-positives</link>
    <description>A labels&#39; court filing shows DistroKid tests uploads with an undisclosed third-party &quot;prediction model&quot; and has made statements about false positives, none of it published for artists. Here is what is known about distributor AI detection, what you control, and how to protect a release that gets misjudged.</description>
    <pubDate>Fri, 23 Oct 2026 04:30:00 GMT</pubDate>
    <dc:creator>We are the music makers</dc:creator>
    <category>workflows</category><category>ai-music</category><category>ai-detection</category><category>distrokid</category><category>distribution</category>
  </item>

  <item>
    <guid>https://wearethemusicmakers.co/news/2026/09/ai-music-detectors-not-proof-artifactbench-benchmark</guid>
    <title>AI music detectors are not proof: what a benchmark of eight of them found</title>
    <link>https://wearethemusicmakers.co/news/2026/09/ai-music-detectors-not-proof-artifactbench-benchmark</link>
    <description>Intrect tested eight public AI music detectors on the same 2,104 files. The public Deezer research model flagged 13% of real tracks and caught 64.6% of AI ones, and two detectors flagged more than two-thirds of real music as AI. Vendors advertise 99%. Here is how to read the gap, and what to do if a score lands on your release.</description>
    <pubDate>Mon, 14 Sep 2026 04:30:00 GMT</pubDate>
    <dc:creator>We are the music makers</dc:creator>
    <category>ai</category><category>ai-music</category><category>ai-detection</category><category>deezer</category><category>benchmarks</category>
  </item>

  <item>
    <guid>https://wearethemusicmakers.co/news/2026/08/aria-beatport-draw-the-ai-line</guid>
    <title>ARIA and Beatport both draw an AI line. Only one lets a human hear the appeal</title>
    <link>https://wearethemusicmakers.co/news/2026/08/aria-beatport-draw-the-ai-line</link>
    <description>Australia&#39;s ARIA Charts now exclude wholly AI-generated recordings, and Beatport withholds fully or majority AI tracks at upload. Both rules sound similar. The difference that matters to artists is who decides, and what happens when they get it wrong.</description>
    <pubDate>Tue, 25 Aug 2026 04:30:00 GMT</pubDate>
    <dc:creator>We are the music makers</dc:creator>
    <category>artist-rights</category><category>aria</category><category>beatport</category><category>ai-detection</category>
  </item>

    </channel>
  </rss>
