IBM watsonx Model Risk Register is a Kinetic Gain platform-signal repo for IBM watsonx. It turns synthetic AI governance, model-risk, lineage, IBM Cloud IAM, and release-readiness signals into a board-readable register.
Which watsonx model use cases need risk, disclosure, or access evidence before executive sign-off?
- Model risk: unclassified assistant use cases, risk tier, approved use boundary, and disclosure posture.
- Evidence: prompt, response, retrieval, and tool-call audit streams.
- Lineage: source dataset cards, retention policy, and approved grounding or training use.
- Access: IBM Cloud IAM drift around watsonx service roles and governance groups.
- Release: eval gates for model updates, retrieval changes, and prompt releases.
- Attestation: system cards, data cards, vendor disclosures, and incident route packets.
npm install
npm run verify
npm start
npm start -- --json# IBM watsonx Model Risk Register
IBM watsonx: watch posture, 74/100 risk, $0 recoverable or protected value.
Board question: Which watsonx model use cases need risk, disclosure, or access evidence before executive sign-off?
Highest-risk lane: Unclassified assistant use case (AI governance)
The repo builds a synthetic board surface into site/index.html.
npm run build:siteThis repo uses synthetic demonstration data only. It does not connect to live IBM watsonx tenants, export customer records, store credentials, or publish admin-console screenshots.
- Company/platform signal: IBM, IBM watsonx, IBM Cloud IAM
- Domain: AI governance and model risk
- Live surface family: https://watsonx.kineticgain.com/
- Apex: https://kineticgain.com/
- Portfolio: https://portfolio.kineticgain.com/