Now
Everything moving in the standards community
Discussions and publisher posts from every source the feed publishes, newest first. Each entry links to its source; nothing here is a copy of what was said.
Curated by Nicholas J. Dugan.
Sources
- FHIR Chat — 111 messages this week across 31 feeds · last message 4 hours ago
- HL7 Mailing List — 3 messages this week across 2 feeds · last message 36 hours ago
- HL7 Blog — 1 post this week · last post 3 days ago
Feed checked 4 minutes ago. Counts are the feed's own, for the last seven days.
Specification: 1 of 88
Sources: All
Channels: All
Stage: All
Impact: All
More
Columns: All
Showing 2 of 2 (73 before filters) · summarised discussions only
| Question | Next step | Specifications | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Patient consent for agent reads: Consent vs scopes (opens chat.fhir.org) | How do implementers identify the agent as the Consent actor (OAuth client vs Device), whether agent grants map to R5 Permission instead of Consent, and how to show patients what the agent read (AuditEvent per read vs coarser)? | Expected: implementers may treat per-request Consent period checks and status-based revocation as the model, but nothing is decided or changed in a specification yet. | No direct answers to the three open questions yet. Replies point to the AI transparency IG (Provenance for AI-influenced resources, nearing publication) and a draft Consent About AI IG using PurposeOfUse codes, with a Device used only for specific AI models. Suggested AuditEvent if Provenance is insufficient. A participant asked whether checking provision.period per request rather than at token issue is expected, and whether the search/include leak is the same problem as OAuth scope leakage; the | Group members are expected to respond to the open questions on agent identity, R5 Permission mapping, and read auditing. | FHIR R4 Consent, FHIR R5 Permission, SMART on FHIR, AI Transparency IG, Consent About AI IG, FHIR AuditEvent, FHIR Provenance | — Cites no HL7 Jira item. | Machine summary · unreviewed · claude-haiku-5-5 · analysed Nov 7, 2026 | Agents on FHIR | FHIR Chat | 8 messages this week · 13 hours ago |
| AI-Transparency require security tag? (opens chat.fhir.org) · chat quiet | Should the AI transparency IG make a security label (meta.security) mandatory to flag AI-created or AI-touched content, or is it better placed in meta.tag or Provenance? | No change to the specification is decided; a proposed SHALL on meta.security for AI-generated content is unresolved, and the value-set review may remove codes bound to meta.security. | No consensus. One participant supports a SHALL on a single meta.security label as a quick signal, with fuller Provenance later. Others argue AI is not purely a security matter and fits meta.tag or Provenance better, and that responsibility and the degree of AI instruction and review are not captured by a binary flag. One participant argues AI assertions are trust-related and belong in meta.security. A participant says the AI level is covered in Provenance and the security label is only a coarse | The work group triages FHIR-58669 on the questionable value-set codes; no decision on the meta.security SHALL is recorded. | FHIR security labels (meta.security), FHIR Provenance, FHIR meta.tag, AI Transparency IG |
| Machine summary · unreviewed · claude-haiku-5-5 · analysed Oct 15, 2026 · inferred | Artificial Intelligence/Machine Learning (AI/ML) | FHIR Chat | chat quietChat quiet since Aug 22 · FHIR-58669 moved 12 days ago |
Links open the source, which asks you to sign in if the discussion needs it. This page shows the newest entries the feed returns, not a complete archive. Summaries, including the Impact sentence, are written by a model from the discussion and say so until they are reviewed; they name no one. Stages, Impact categories and Jira statuses come from the thread and from HL7 Jira, not from a work group's formal record. A discussion whose chat has gone quiet stays listed while its Jira items are open or moving.