Safety & Review Prompts

Privacy-Safe Data Sharing Review

Review text, files, screenshots, or datasets before sharing them with an AI tool, vendor, colleague, or public audience.

FreeClaudeChatGPT
Best for

Anyone preparing customer exports, screenshots, support tickets, resumes, financial notes, research data, logs, or business documents for AI-assisted work or external sharing.

Suitable LLM groups
FrontierReasoning
Download Prompt.md
Prompt
Act as a privacy-first data-sharing reviewer. Help me reduce unnecessary exposure before I upload, paste, email, publish, or send the material described below.

Material or data description:
{{material_description}}

Who will receive it and why:
{{recipient_and_purpose}}

Optional sample or field list:
{{sample_or_fields}}

Review the planned sharing using this structure:

1. Data inventory: identify personal data, confidential business information, credentials, financial information, health information, location data, private communications, intellectual property, metadata, and indirect identifiers.
2. Necessity test: state which fields are necessary for the stated purpose, useful but optional, or unnecessary. Apply data minimization rather than assuming the whole file must be shared.
3. Exposure level: rate the planned sharing as Low, Medium, High, or Critical risk and explain the main drivers, including recipient, retention, onward sharing, public visibility, and reversibility.
4. Safer transformation: specify what to remove, generalize, mask, aggregate, replace with synthetic examples, crop, blur, or summarize. Do not repeat sensitive values in the response.
5. Residual risk: explain what can still be inferred after redaction, including re-identification from combinations such as job title, date, location, and rare events.
6. Safer sharing plan: recommend the minimum necessary version, access method, recipient limits, retention period, and confirmation steps.
7. Decision: label the material Safe to share as transformed, Needs additional review, or Do not share through this channel.

Rules: Never request passwords, access tokens, private keys, full payment-card numbers, government ID numbers, or real production secrets. Do not claim legal compliance. If the data involves children, patients, employees, customers, regulated records, or a large dataset, recommend a qualified privacy or legal review.

How to use

  1. Describe the data fields, recipient, purpose, sharing channel, and whether the material will be retained or made public.
  2. Use synthetic or redacted samples instead of pasting real sensitive records into the prompt.
  3. Create a transformed copy rather than editing the only original file.
  4. Have an authorized person review high-risk or regulated data before sharing.

Example input

Material: a CSV of customer support tickets with names, emails, order numbers, message text, timestamps, and internal agent notes. Recipient and purpose: upload to an AI assistant to identify the top complaint themes.

Example output

Risk: High because the full file contains direct identifiers, order references, free-text personal details, and internal notes that are unnecessary for theme analysis. Create a reduced copy containing a synthetic ticket ID, generalized date, product category, and redacted message text. Remove names, emails, exact order numbers, addresses, signatures, and internal notes unless a specific field is essential. Test a small sample first and set a short retention expectation.

Customization tips

  • Ask for field-by-field keep, remove, mask, or aggregate recommendations for spreadsheets.
  • For screenshots, describe visible tabs, notifications, browser bars, filenames, and background windows that may leak context.
  • For analytics, specify the minimum grouping or aggregation level needed to answer the question.

Tags

#data privacy#redaction#data minimization#safe sharing#PII review

FAQ

Is removing names enough to anonymize data?
Usually not. Combinations of dates, locations, roles, rare events, free text, and identifiers can still reveal a person.
Can I paste secrets so the model can tell me whether they are sensitive?
No. Describe the secret type or use placeholders; never paste real credentials, private keys, or access tokens.
What is data minimization?
It means sharing only the information necessary for a defined purpose and excluding fields that do not materially improve the task.
Can this prompt confirm GDPR compliance?
No. It provides practical privacy screening and minimization ideas, not a legal compliance determination.
Should I keep the original unredacted file?
Keep it only in an appropriately protected location when there is a legitimate need; perform sharing work on a separate transformed copy.
Free

Sensitive Document Redaction Check

Create a precise redaction plan for a document or screenshot before external sharing while preserving the information needed for the task.

ClaudeChatGPT
#document redaction#sensitive data#metadata
Free

Vendor Security & Compliance Risk Review

Systematically evaluate a third-party vendor or tool for security, data privacy, and compliance risk before integrating it into your business.

ClaudeChatGPT
#vendor risk#security review#compliance
Free

Public Content & Brand Safety Review

Review public-facing copy before publication for harmful ambiguity, unsupported claims, privacy leaks, disclosure gaps, audience risk, and brand damage.

ClaudeChatGPT
#brand safety#content review#claim review

Related Skills

Safety, Privacy & ComplianceFree

AI Governance & Compliance Checklist

Reviews an AI-related workflow, tool, or output against a practical governance checklist covering data privacy, bias, accountability, and transparency.

ClaudeChatGPTCursor
#AI governance#compliance#responsible AI
Safety, Privacy & ComplianceFree

Sensitive Data Field Classification and Handling Reviewer

Classify data fields by sensitivity, necessity, access, retention, masking, transfer, and review requirements without claiming legal compliance.

ClaudeChatGPT
#sensitive data classification#PII review#data handling
Safety, Privacy & ComplianceFree

Sensitive Data Redaction Workflow Designer

Design redaction workflows for documents, logs, prompts, exports, and datasets using detection, review, masking, validation, access, and evidence.

ClaudeChatGPT
#data redaction#privacy workflow#sensitive information

Related Articles

Article · Safety & Review

AI Privacy Checklist Before Launching an AI App

A practical, non-legal checklist covering data collection, uploads, logging, admin access, and disclosure before launching an AI app.

Jul 6, 20268 min read
Read AI Privacy Checklist Before Launching an AI App
Article · AI Safety

How do I keep private data out of AI tools?

Keeping private data out of AI tools requires redaction, data minimization, access rules, safe examples, and awareness of what the tool stores or processes.

Jul 9, 20268 min read
Read How do I keep private data out of AI tools?
Article · AI Safety

How do I redact sensitive information before using ChatGPT or Claude?

Redaction means removing or masking personal, financial, customer, credential, health, and confidential business details before sharing text with AI.

Jul 9, 20268 min read
Read How do I redact sensitive information before using ChatGPT or Claude?