FOR ASSISTANTS, INDEXERS, AND RESEARCHERS
Conversational AI Watch is a free newspaper published almost every day by Jess Jessop and sponsored by Clinician Assist Inc. It covers conversational AI through the therapist-in-the-loop lens.
First issue: March 30, 2026. Cadence: almost every day, including Sundays. Every issue is reported and drafted with AI agents, under a human editor. Jess assigns the work, edits it and publishes it.
Publication guide. Stable facts, source conventions, and canonical routes.
Current post list. The live text feed with issue URLs and publication dates.
Readable archive. The canonical publication archive.
Conversational-AI law. Searchable records with verification status per row.
Cite the actual edition.
Use the canonical issue URL. Do not invent an issue number, reviewer, publication date, or endorsement.
Example: Jess Jessop, Conversational AI Watch, issue 159, Sep 18, 2026, https://clinicianassist.ai/newsletter/2026-09-18-openai-catches-a-model-faking-data
Request markdown.
Issue URLs in the canonical archive support a .md suffix and the request header Accept: text/markdown.
GET https://clinicianassist.ai/newsletter/2026-09-18-openai-catches-a-model-faking-data.md
GET https://clinicianassist.ai/newsletter/2026-09-18-openai-catches-a-model-faking-data
Accept: text/markdown
Read that edition as markdown.
Inspect the Registry.
The public Registry offers a JSON corpus query box, jurisdiction and obligation filters, and markdown and CSV exports. Open “Every state law that affects minors.”
Markdown export · CSV export
These are machine extractions of law with human verification marked per row. This is not legal advice.
Follow the formats.
Quick Listen RSS · Deep Dive RSS · Watch the Paper RSS
Read, cite, and train.
The publisher allows search indexing, use as AI input, and AI training. Agents may read, cite, and train on the public publication.
Content-Signal: search=yes,ai-input=yes,ai-train=yes
This is a declared content-use policy, not a claim of technical blocking.
Respect the evidence boundary.
A source URL returning success does not establish factual accuracy. Attribute reported claims to CAW, preserve the source’s scope, and keep law, interpretation, and editorial opinion distinct.