Tidal AI Policy

Original Article ↗ Hacker News Discussion ↗

Tidal has introduced an AI music policy that allows AI-generated tracks on the platform but labels them, blocks them from earning royalties, and promises to remove impersonations and obvious spam. Commenters broadly welcome efforts to curb “AI slop” and protect human artists’ income, but raise concerns about vague definitions of “AI-generated,” unreliable detection technology, and the risk that platforms will quietly promote royalty‑free AI music to boost margins. Many users say they want a simple way to hide AI content entirely, while others argue that AI can be a legitimate creative tool and that the real problem is incentives and curation, not the technology itself.

Overview of Tidal’s AI Policy

  • Tidal will:
    • Accept AI-generated music.
    • Tag tracks it detects as wholly AI-generated with an “AI” badge.
    • Disallow monetization and direct sales for 100% AI-generated tracks.
    • Remove impersonations and obvious fraud, and may extend “substantially AI-generated” rules as detection improves.
  • Many commenters see this as a pragmatic middle ground: accept that AI exists, constrain abuse, and protect payouts for human-created work.
  • Others think merely allowing AI music is “another one bites the dust” and want outright bans or at least a global “hide AI” toggle.

Monetization, Incentives, and Platform Economics

  • Supporters: Turning off royalties for AI content removes the main incentive for spammy “AI slop” floods.
  • Critics:
    • Tidal still profits from streams of AI tracks while paying no royalties, creating a strong incentive to recommend AI music in playlists.
    • Some argue Tidal should then refuse AI uploads entirely, rather than keep the revenue.
  • Broader concern that streaming platforms will favor free-to-them AI tracks over human artists in algorithmic discovery.

Definitions, Detection, and Edge Cases

  • “AI-generated” is defined as works wholly or substantially made by generative AI with minimal human creative input, but what counts as “substantial” is unclear.
  • Edge cases: human lyrics + AI backing, AI vocals over human instrumentals, AI stems in otherwise human productions.
  • Detection is widely viewed as hard and error-prone, especially as AI improves and producers mix AI samples into conventional workflows.
  • Fear of false positives harming experimental electronic/spectral music and sample-heavy genres like EDM and hip hop.

Copyright and Legal Ambiguity

  • Some point to U.S. rulings that raw generative output lacks copyright and is effectively public domain.
  • Others note this varies by jurisdiction and may hinge on “skill and judgment” or level of human control.
  • Debate over whether a prompt-writer deserves royalties when the model was trained on others’ copyrighted works.

Curation, User Choice, and Value of AI Music

  • Many users report “AI slop” flooding discovery feeds on Tidal, Spotify, YouTube, and ebook platforms, and want better curation or hard filters for AI.
  • Some enjoy AI music as background or mood (“lofi beats,” synthwave, EDM, focus tracks) and see it as just another tool.
  • Others see AI-generated works—especially one-shot prompts—as hollow, derivative, or ethically tainted, arguing that music’s value comes from human experience and expression.