Quiz: Laying the Foundations for AI at Work

Pass mark: 70%. You can retry as many times as you need.

1 Your colleague says: "AI will replace my job, so there's no point learning it." Which response is the most accurate? Single answer That's right: you should wait until the situation is clear before doing anything. AI affects tasks more than whole jobs, so the useful step is to identify which of your tasks it can help with and how to check it. AI cannot do anything useful, because it makes mistakes. Only technical jobs are affected, so office workers can ignore it.

2 Which of these tasks are good candidates for a first AI use case? (several answers) Multiple answers possible Rewriting routine reminder emails in a friendly tone, which you re-read before sending Deciding which candidates to shortlist from their CVs Turning your own rough meeting notes into a structured summary using an approved tool Certifying that a gas installation meets safety standards

3 Léa has fifteen tasks scored on repetitiveness, textual nature, tolerance to error and data sensitivity. One task, drafting contract clauses for clients, scores high on repetitiveness and text but low on tolerance to error and high on sensitivity. What is the best decision? Single answer Start with it, since it is repetitive and textual. Put it first because it has the highest potential value. Delegate it entirely to the AI to save the most time. Keep it out of the first three use cases; revisit it later with strong safeguards and professional review.

4 Karim, a freelance consultant, wants to paste a client's contract into a free personal chatbot account to get a summary. What should he check or do first? (several answers) Multiple answers possible Whether the tool retains the data and may reuse it for training Whether his confidentiality obligations to the client allow it at all Whether the summary will be shorter than the contract Whether he can anonymize the document or use a tool that is suitable for confidential material

5 Your team keeps writing similar prompts from scratch and results vary a lot between colleagues. What is the best way to improve consistency? Single answer Ask everyone to memorize the best prompt. Build a shared, versioned prompt library with allowed data, checks and known limits for each prompt. Ban the tool until everyone writes prompts in the same way. Let each person keep prompts in their own chat history.