Evaluate task safety and privacy for AI drafting

Assess whether a specific task or text draft is safe to process with an AI model based on privacy, accuracy, and professional stakes.

You getRisk Rating, data Privacy Check and hallucination Hazard

Prompt

Prompt
Act as an AI risk and privacy advisor. Evaluate the described task or pasted text to assess whether it is safe for direct AI drafting.

- Identify any private, regulated, proprietary, or sensitive personal data present.
- Assess the operational risk of subtle factual hallucinations or technical errors.
- Classify the task safety level into Low Risk, Moderate Risk, or High Risk.
- Provide concrete guidelines on what parts to anonymize or double-check manually.
- Plain notice: This is general operational information only, not legal or formal compliance advice; consult your organization's legal or security officer for company policies.

Output format:
- Risk Rating
- Data Privacy Check
- Hallucination Hazard
- Safe Usage Recommendations

Here is the task description or text to evaluate:

Example output

Example for: pasting customer medical refund requests into an AI assistant

Risk Rating

High Risk

Data Privacy Check

The sample text includes customer names, invoice identifiers, and medical procedure descriptions. Submitting this directly into general public AI tools violates consumer privacy principles and health data compliance standards (such as HIPAA). Personal identifiable details must be completely scrubbed before processing.

Hallucination Hazard

Moderate to High. If the model is asked to summarize refund eligibility rules or dispute amounts, small arithmetic misreadings or policy errors can lead to incorrect financial outcomes.

Safe Usage Recommendations
  • Replace all actual names, customer account numbers, and clinic names with generic tokens like Customer 1 or Clinic A.
  • Do not allow the AI to make binding approval decisions; use it solely to rewrite customer service explanations from approved templates.
  • Ensure a human reviewer compares all stated figures against official billing software records.

Notice: General operational information only, not legal or formal compliance advice; consult your organization's legal or security officer for company policies.

Written by AI from this prompt. Your results will vary.

Paste your proposed prompt, task description, or document excerpt into ChatGPT, Claude, or Gemini. Follow up by asking how to sanitize the text to lower its risk category.

Make it repeatable

Keep using this prompt without pasting it in every time.

Keep exploring

More in AI Tools & Prompt Engineering
AI Tools & Prompt Engineering

Enforce strict refusal rules for missing context

Generate airtight prompt guardrails that stop AI models from guessing when source facts are absent. Use this whenever hallucinated answers threaten critical team workflows.