From the Standards Committee
The audio economy is being rewritten. We publish analysis, context, and strategy for the creators, broadcasters, and researchers caught in the middle of it.
What No One Is Saying About Colorado's ADMT Rules
Colorado's new AI rules require training data documentation. But nobody in the public comment record asked the obvious question: how does anyone verify the documentation is accurate? The insurance industry already solved this problem. Here is the argument nobody else made.
September 24, 2026The Verification Gap: Why AI Training Data Documentation Needs Independent Certification
Every new AI regulation requires developers to document their training data. None of them explain how anyone checks whether the documentation is true. This is not a new problem — and every industry that has faced it solved it the same way.
September 24, 2026Audio Data Is the Next AI Liability Frontier
Text and images dominated the first wave of AI training data lawsuits. Audio is next — and the liability exposure is worse. Voice carries biometric identity, copyright entanglement, and consent requirements that text never did. The insurance market has already priced the risk.
September 24, 2026What ISO CG 40 47 and CG 40 48 Mean for AI Companies
In January 2026, Verisk ISO released endorsements that strip AI liability from standard commercial insurance. 95% of carriers are adopting them. If you build, deploy, or sell AI systems, your insurance coverage probably changed this year.
September 24, 2026HITRUST for AI Training Data: Why the Healthcare Precedent Matters
Healthcare solved the third-party data verification problem with HITRUST. Over 81% of hospitals and 83% of health plans use it. AI training data has the same problem. The same architecture works — and it already exists.
September 24, 2026The Audio Data Gap: Why Indie Broadcasters Are Sitting on a Gold Mine
The music industry has SoundExchange. Union actors have SAG-AFTRA. Independent podcasters, community radio stations, and faith-based broadcasters have nothing. Their archives contain exactly what AI companies need most, and right now, nobody is protecting it.
September 12, 2026Your Voice Is Not Training Data (Unless You Say So)
Every podcast episode, voice reel, and broadcast segment you have ever published is a potential training input for AI models. That is not a hypothetical. It is happening right now, without your knowledge or consent. Here is what you can do about it.
September 12, 2026What a Member-Owned Audio Data Standard Looks Like
A standards body that accepts funding from the companies it regulates is a standards body that will eventually serve those companies. We designed Box Commons to make that structurally impossible. Here is how the governance model works and why it matters.
September 12, 2026Get Your Station Protected
Join the growing coalition of independent broadcasters securing their web streams and archives against unauthorized AI extraction.