Privacy & Safety

AI Foundations: What It Is and How to Use It Beginner · 36 min How to Create an AI Use Policy for Your Small Team Beginner · 19 min Know When to Trust What AI Tells You Beginner · 20 min

Before You Write a Single AI Rule How to Create an AI Use Policy for Your Small Team · 2 min Build Your Personal Evaluation Protocol Know When to Trust What AI Tells You · 3 min Checking the Work AI Foundations: What It Is and How to Use It · 5 min Finish It and Share It How to Create an AI Use Policy for Your Small Team · 4 min How to Brief It AI Foundations: What It Is and How to Use It · 6 min Introduction Know When to Trust What AI Tells You · 3 min Keeping It Safe: Data, Money and Reputation AI Agents for Business Owners: What They Are and What to Hand Them · 8 min Know What You’re Looking For Know When to Trust What AI Tells You · 4 min Make AI Show Its Work Know When to Trust What AI Tells You · 3 min Next Steps Know When to Trust What AI Tells You · 1 min Push Until AI Disagrees With Itself Know When to Trust What AI Tells You · 3 min Rapport, Not Manipulation Emotionally Intelligent AI · 8 min

Anonymisation Removing identifying details so data can be used safely. Bias Skew a model inherited from the material it learned from. Citation A reference to a source — which AI can invent wholesale. Commercial Use Whether you are allowed to use AI output in paid work. Copyright Who owns AI-generated work — a genuinely unsettled question. Data Retention How long a provider keeps what you send it. Deepfake Synthetic audio or video made to look like a real person. Fact-Checking Verifying specific claims outside the tool that produced them. Guardrails The limits built into an AI tool to stop it doing certain things. Hallucination When an AI states something false as if it were true. Human in the Loop A person who checks the output before it counts. Jailbreak Wording a request to get round a tool's restrictions. Knowledge Cutoff The date after which a model knows nothing, unless it can look things up. Licence The terms you agreed to when you started using a tool. Open-Source Model A model whose weights anyone can download and run. Personal Data (PII) Information that identifies a living person — and must not be pasted casually. Training Data The material a model learned from — and the limit of what it knows. Transparency Being open about where and how you use AI. Watermark A marker showing content was machine-generated.

Fact-check an AI answer Make the model argue against itself before you trust the output. Draft an AI use policy for your team A first draft your team will actually follow, not a compliance document nobody reads. Is AI the right tool for this job? A short check before you spend an afternoon automating something that did not need it. Check a draft for anything that must not be published The last read before something goes out: claims you cannot evidence, and anything that is not yours to say. Is this a job for an AI agent? Score any task against four tests and find out whether it suits an agent, an automation, or you. Find the jobs in my week an agent could take on A short interview about your week that ends in a ranked shortlist of agent jobs, and a list of things that should stay with you. Write my Agent Brief Turns a job you want to delegate into an eight-part brief an agent can actually follow, including when it must stop and ask. Stress-test my Agent Brief before it runs Finds the gaps, guesses and risks in your agent instructions before the agent finds them for you. Work out why an agent run went wrong Traces a bad result back to its cause (brain, tools, instructions or memory) and suggests the one change to make before the next run. Review my agent plan and write my next brief Hand your completed Delegation Plan workbook to AI. It checks your plan for risk, writes the Agent Brief for your next job and turns your 90 days into a weekly checklist.