Lesson 3 of 4 · about 5 min
Guardrails for project managers
When you finish, you can decide what never goes into an AI tool, how to keep reports honest, and how to handle risks about people and data.
Stuck? Ask the AI tutorProject information can be sensitive
Project documents often hold budgets, contracts, supplier terms, unreleased plans, security details and notes about people’s performance. Contracts and non-disclosure agreements may forbid sharing them. Some projects, such as in healthcare, finance or government, involve regulated information.
Do not paste confidential plans, contract text, credentials, personal data about team members or customers, or security details into a consumer AI tool unless your employer has approved that tool for that data. Replace names with roles, remove figures, and describe things generally.
Accuracy and honest reporting
Tools can create tidy but wrong plans: a task order that ignores a real dependency, a timeline that looks possible but is not, or a status that sounds cheerful. An optimistic report that hides a problem costs far more later than a plain one now.
Check every date, owner, dependency and number. Treat a generated schedule as a conversation starter. Keep your sources of truth, such as your tracking tool and the team’s own words, as the real record, and update them yourself.
Fairness and people decisions
Be careful about using AI to assess people: ranking team members, judging performance from chat logs, or deciding who is overloaded. Tools can be biased, and the data they see is partial. These decisions need your direct conversation and your organization’s process.
Also watch for subtle bias in who gets which tasks when you ask a tool to suggest assignments. It may reflect stereotypes. Assign based on skills, interest and capacity you have discussed with the person.
Disclosure and saying no
If a document that matters, like a plan sent to a client, was drafted with AI help, follow your organization’s guidance on saying so. At minimum, be ready to explain how you checked it. Some contracts and industries have rules on using automated tools.
You can say no to risky shortcuts: putting the client’s confidential spec into a public chatbot, auto-sending status reports without review, or using AI to monitor team members in ways they do not know about. A good reply is, “I want to protect the project and the team. Can we use an approved tool or remove the sensitive parts first?”
Practice now
- List the kinds of information in your current or past project that should never go into a consumer AI tool. Compare with your organization's policy.
- Take a plan or schedule a tool produced and find three assumptions or dependencies that are wrong or missing.
Use any AI assistant you have. Do not paste private, client or patient information into it.
Check yourself
1. A project manager pastes a client's confidential specification into a free chatbot to summarize it. What is the main issue?
2. Why should you check dates and dependencies in a generated plan?
3. Which use of AI about people should you avoid?
Ask the AI tutor
Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.
AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.