Security: What NOT to Put in AI Tools
AI Security for Administrative Professionals
As an administrative professional, you handle sensitive information daily — employee records, financial data, client information, strategic plans. Using AI tools requires understanding what is safe to share and what is not.
🚫 NEVER Put These in AI Tools
- Personally Identifiable Information (PII): Social Security numbers, passport numbers, driver's license numbers, dates of birth, home addresses
- Financial data: Bank account numbers, credit card numbers, payroll details, tax identification numbers
- Health information: Any medical records, insurance claims, health-related data (protected under HIPAA in the US)
- Confidential company data: Trade secrets, proprietary processes, unreleased financial results, strategic plans
- Client/customer information: Customer lists, client contact details, client financial data (unless you have explicit authorization)
- Legal documents: Contracts under negotiation, litigation details, attorney-client communications
- Authentication data: Passwords, API keys, security codes, access credentials
- Employee records: Performance reviews, salary information, disciplinary records, background check results
✅ Generally Safe to Share with AI
- Generic, non-confidential text for drafting (e.g., "Write a memo about a new parking policy")
- Publicly available information (already on your website, in press releases)
- Generic templates and formats without real data
- Meeting agendas and summaries that don't contain sensitive decisions
- General questions about how to use software or tools
- Anonymized data (with names, numbers, and identifiers removed)
The Anonymization Technique
When you need AI help with a document that contains sensitive information, anonymize it first:
| Original (Don't Share) | Anonymized (Safe to Share) |
|---|---|
| "John Smith's salary is $85,000" | "Employee A's salary is $X" |
| "Acme Corp, 123 Main St, contract value $500K" | "Client X, contract value $Y" |
| "SSN: 123-45-6789" | [Remove entirely] |
Understanding AI Data Policies
Different AI tools handle your data differently:
| Tool | Free Tier Data Use | Paid/Enterprise |
|---|---|---|
| ChatGPT | May use chats to train models (opt-out available in settings) | Team/Enterprise: data not used for training |
| Copilot | Commercial data protection for eligible accounts | M365 Copilot: data stays within tenant |
| Gemini | May use conversations to improve services (opt-out available) | Workspace: data not used for training |
⚠️ Key actions to take today:
- Check your AI tool's privacy settings — turn off data sharing/training if available
- Ask your IT department if your company has an AI usage policy or approved tools
- Use enterprise versions when available — they offer better data protection
- When in doubt, leave it out — if you're unsure whether something is safe to share, don't share it
Company AI Policies
Many organizations are developing AI usage policies. As an administrative professional:
- Follow your company's policy — if one exists, it overrides any general advice
- Ask if you're unsure — check with IT or compliance before using AI for sensitive tasks
- Advocate for clear guidelines — if your company doesn't have an AI policy, suggest they create one
- Use approved tools only — your company may have approved specific AI platforms over others
Critical security guidelines for administrative professionals using AI. Learn what information is safe and what must never be shared with AI tools.
Rating
0
0
There are no comments for now.
Join this Course
to be the first to leave a comment.