Learn

Using AI safely at work

Short, sourced answers for teams that already use ChatGPT and Claude with client data, and the kits to act on them.
Find your answerGet a kit

Start with your job

What do you need to do?

Choose the job in front of you. Each path opens the matching questions.

Check a vendor or an AI tool

  • How do you tell real AI governance from theater?
  • What can and cannot be guaranteed about an AI system?
  • Who built the AI model, and on what?

Build or run AI agents safely

  • Is it safe to let an AI agent act on company systems?
  • What is the OWASP Top 10 for LLM applications?
  • What is prompt injection, and how do you defend against it?

Understand the terms

  • When people say "AI," what are they talking about?
  • What is the difference between training and inference?
  • AI glossary: the terms that carry weight

By profession

Guidance for the work you do

Using AI safely at work

Practical answers for teams using ChatGPT, Claude, and other AI tools with client and company data.

Browse questionsWhat happens to what you type into an AI chat? · What are the security risks of generative AI at work? · ChatGPT data leakage: how do you stop company data leaking into AI tools?9
  1. 01What happens to what you type into an AI chat?
  2. 02What are the security risks of generative AI at work?
  3. 03ChatGPT data leakage: how do you stop company data leaking into AI tools?
  4. 04Is there a private ChatGPT for business?
  5. 05Should a company ban AI at work?
  6. 06Where does your data go when people use AI?
  7. 07Is ChatGPT safe to use at work?
  8. 08What is shadow AI, and how do you detect it?
  9. 09What is PII data, and what counts as PII?

Rules and compliance

Policies, healthcare requirements, enforceable controls, liability, and rules that cross jurisdictions.

Browse questionsAI acceptable use policy: a copyable template · Is ChatGPT HIPAA compliant? · Is Claude HIPAA compliant?7
  1. 10AI acceptable use policy: a copyable template
  2. 11Is ChatGPT HIPAA compliant?
  3. 12Is Claude HIPAA compliant?
  4. 13Is AI HIPAA compliant?
  5. 14Are your AI controls enforced, or just written down?
  6. 15Who is liable when an AI system causes harm, or an agent acts?
  7. 16Can one AI system comply with several regulatory regimes at once?

How AI governance works

How to assess governance claims, enforce rules while AI runs, and keep records that people can inspect.

Browse questionsWhat is the OWASP Top 10 for LLM applications? · What is prompt injection, and how do you defend against it? · Is it safe to let an AI agent act on company systems?11
  1. 17What is the OWASP Top 10 for LLM applications?
  2. 18What is prompt injection, and how do you defend against it?
  3. 19Is it safe to let an AI agent act on company systems?
  4. 20What should an AI governance framework include?
  5. 21How do you tell real AI governance from theater?
  6. 22Who built the AI model, and on what?
  7. 23What does AI enforcement at runtime actually look like?
  8. 24What can you see while your AI is running?
  9. 25Can you prove afterwards who did what with AI?
  10. 26What can and cannot be guaranteed about an AI system?
  11. 27What can an AI governance layer not do?

How AI works

What the word covers, what a model is made of, and why running AI costs more than building it.

Browse questionsWhen people say "AI," what are they talking about? · Which AI systems can have their logic audited, and which cannot? · Is it predicting something, or making something?8
  1. 28When people say "AI," what are they talking about?
  2. 29Which AI systems can have their logic audited, and which cannot?
  3. 30Is it predicting something, or making something?
  4. 31Does the AI advise a person, or act on its own?
  5. 32What is a model weight, and what is a transformer?
  6. 33What is the difference between training and inference?
  7. 34Does all AI need a hyperscale data center?
  8. 35What is the difference between open weights and open source?

Reference

The terms that carry weight, and the AI statistics that do not survive a trip to their source.

Browse questionsAI glossary: the terms that carry weight · Which AI statistics should you stop repeating?2
  1. 36AI glossary: the terms that carry weight
  2. 37Which AI statistics should you stop repeating?

Free kits

Turn an answer into a working practice

Preview each kit on the page that owns its search. The email step stays with the download.

HIPAA and AI checklist

A printable verification list for teams considering ChatGPT or Claude with protected health information.

Who is liable when AI acts

A responsibility map for one workflow, with the rule for lawyers, tax preparers, and healthcare practices at each stage.

AI numbers to stop using

Seven repeated figures against their primary sources, with the sentence to use instead and a worksheet for the next one.

AI vendor claims, decoded

Fourteen common vendor claims, what each can mean, the proof to request, and the answer that should end the call.

AI terms that matter

Twenty-seven contract and policy terms, each with the question to ask and where it appears in the paperwork.

Videos

Watch AI governance in practice

Watch Sentinel and PII protection in practice. Each video keeps its existing watch page.

All videos

SUPERWISE Sentinel

Put a control in the path

See the documented coverage and limits on the Sentinel coverage statement.

Get SentinelWhat it installs

After installing, sign in or create a free account, then approve this computer.

Windows 10 and 11 · Free for 30 days · Covers Claude Code and Claude Desktop. The browser extension is in preview for ChatGPT and Claude in Chrome and Edge.

Compare Sentinel