On 29 July 2026 the National Institute of Standards and Technology released an initial public draft titled “Guidance and Templates for Public-Facing AI Documentation: An AI Standards ‘Zero Draft’”. NIST stated that it will consider input received by 16 September 2026 for the subsequent, and possibly final, revision. The draft carries the identifier NIST AI 300-1 ipd.
The publication is part of NIST’s AI Standards Zero Drafts pilot. That project is deliberately unusual in process terms: rather than waiting for a standards development organisation to begin work, NIST collects input on topics that already have a science-backed body of work behind them and drafts a preliminary, stakeholder-shaped text - a “zero draft” - that a formal standards body can then take up. The aim is to broaden participation and shorten the time it takes for a standard to exist at all.
The subject of this particular draft is documentation intended for public audiences rather than internal engineering artefacts. That is a meaningful distinction. Internal model documentation has existed in various forms for years, but there has been no common template for what a developer should publish about a system so that customers, auditors and regulators can compare one system against another on consistent terms.
For a business leader, the practical value is anticipatory. Disclosure obligations are arriving in parallel from several directions - EU AI Act transparency duties, US state frontier-model laws, and procurement requirements. A common NIST template, if it holds through the comment period, is the most likely candidate for the artefact all of those regimes end up pointing at. Organisations that already produce structured model or system documentation should read the draft during the comment window and check how far their existing format is from it.