Leila Fielding
Use this when: reviewing a proposed AI use case across multiple risk areas and owners—not making a legal compliance or go-live decision.
WHAT YOU'LL ACHIEVE
Create a working risk picture, record mitigations and identify questions that need accountable human review.
IMPORTANT: A PLANNING AID, NOT CLEARANCE
This browser-local assessment does not determine legal compliance, certify a system, provide definitive risk clearance, or make go-live decisions. Critical gates cannot be overridden here. Use qualified people to review evidence and applicable obligations.
Do not enter personal, confidential or security-sensitive details. Nothing is stored or sent by this tool.
01 / DEFINE THE USE CASE
Context and people affected
02 / RATE WITH EVIDENCE
Assessment areas
Ratings are your provisional view, not model output. “Unknown” is a valid state: record what is missing instead of guessing.
03 / TRACK ACTIONS
Risk register
No register items yet. Add one to record a risk statement, controls and accountable owner.
04 / REVIEW THE REGISTER
Likelihood × impact matrix
Each cell shows register items at the selected likelihood × impact. Unknown ratings stay outside the matrix and must be resolved.
| Impact ↓ / Likelihood → | Low | Moderate | High | Critical |
|---|---|---|---|---|
| Critical | ||||
| High | ||||
| Moderate | ||||
| Low |
PRELIMINARY SUMMARY
Critical gates unresolved, do not treat this as cleared
0 assessment areas or register items rated High or Critical; 10 assessment areas and 0 register items still have unknown ratings.
Non-overridable critical gates
- Use-case purpose is not described.
- Affected people or groups are not described.
- Intended decisions or actions are not described.
- Data sensitivity and privacy classification is unknown or not mapped.
- Data sensitivity & privacy needs evidence or specialist review (unknown).
- Impact & severity needs evidence or specialist review (unknown).
- Security & prompt injection needs evidence or specialist review (unknown).
- Human oversight needs evidence or specialist review (unknown).
- Regulatory & compliance considerations needs evidence or specialist review (unknown).
- Monitoring & response needs evidence or specialist review (unknown).
These gates cannot be dismissed or overridden in this tool.
Flagged unknowns
- Data sensitivity & privacy
- Impact & severity
- Accuracy & reliability
- Bias & fairness
- Security & prompt injection
- Transparency
- Human oversight
- Regulatory & compliance considerations
- Operational dependencies
- Monitoring & response
No overall result here is a go-live decision. A qualified owner must review the evidence, actions and residual risk.
05 / TAKE IT TO A HUMAN REVIEW
Working report
Preview the Markdown report
# AI Risk Assessment: working draft > Planning aid only. This is not legal advice, a compliance determination, certification, definitive risk clearance or a go-live decision. All entries remain in this browser session; no information is sent or saved by this tool. ## Use-case context - Purpose: Not provided - Affected people: Not provided - Intended decisions/actions: Not provided - Data sensitivity: Unknown ## Overall summary - Status: Critical gates unresolved, do not treat this as cleared - Rationale: 0 assessment areas or register items rated High or Critical; 10 assessment areas and 0 register items still have unknown ratings. - Non-overridable critical gates: - Use-case purpose is not described. - Affected people or groups are not described. - Intended decisions or actions are not described. - Data sensitivity and privacy classification is unknown or not mapped. - Data sensitivity & privacy needs evidence or specialist review (unknown). - Impact & severity needs evidence or specialist review (unknown). - Security & prompt injection needs evidence or specialist review (unknown). - Human oversight needs evidence or specialist review (unknown). - Regulatory & compliance considerations needs evidence or specialist review (unknown). - Monitoring & response needs evidence or specialist review (unknown). - Flagged unknowns: - Data sensitivity & privacy - Impact & severity - Accuracy & reliability - Bias & fairness - Security & prompt injection - Transparency - Human oversight - Regulatory & compliance considerations - Operational dependencies - Monitoring & response ## Assessment evidence ### Data sensitivity & privacy - Rating: Unknown - Evidence / unknowns: Not provided ### Impact & severity - Rating: Unknown - Evidence / unknowns: Not provided ### Accuracy & reliability - Rating: Unknown - Evidence / unknowns: Not provided ### Bias & fairness - Rating: Unknown - Evidence / unknowns: Not provided ### Security & prompt injection - Rating: Unknown - Evidence / unknowns: Not provided ### Transparency - Rating: Unknown - Evidence / unknowns: Not provided ### Human oversight - Rating: Unknown - Evidence / unknowns: Not provided ### Regulatory & compliance considerations - Rating: Unknown - Evidence / unknowns: Not provided ### Operational dependencies - Rating: Unknown - Evidence / unknowns: Not provided ### Monitoring & response - Rating: Unknown - Evidence / unknowns: Not provided ## Risk register No register items added. ## Review boundary A qualified owner and relevant specialists must review evidence, residual risks and legal or sector-specific obligations. Critical gates cannot be overridden in this tool.
Review regulatory questions with qualified specialists. This tool does not claim compliance, certification, definitive clearance or safe deployment.