Score generative-AI risk before it becomes audit evidence.
Use this free Gen-AI Risk Scorecard to assess hallucination, prompt-injection, data-leakage, vendor-dependency, and human-oversight exposure for a generative-AI use case. Start standalone or attach the result to your browser-local AI Register.
The output is a board-ready Markdown summary with risk band, dimension scores, and recommended controls.
Nothing is uploaded. This tool runs in your browser, uses the shared browser-local AI Register when you choose a saved system, and stores scorecard results only in localStorage if you click Save.
Five dimensions, one clear readout.
Scores are structured heuristics, not regulatory definitions. Use the result to decide whether a use case needs governance readiness, LLM security testing, vendor review, or a narrower rollout.
Dimension scores
Recommended controls
Need the score turned into an evidence pack?
If this score lands high or moderate, DSE can scope the governance controls, LLM security testing, vendor review, and handoff evidence needed before wider rollout.
Scope a call →This scorecard is a readiness aid and structured heuristic. It is not legal advice, a model validation, a regulatory classification, certification, audit, or guarantee of compliance. Review outputs with your counsel and risk owners before adoption.