Applied AI at work / Module 2
Module 2. Safe use and confidentiality: what never goes into a prompt
Your role: General (any role). The exercises open on its variant. Change role
Time: 3 hours (20 minutes of video lessons, 110 minutes of exercise, 20 minutes of quiz). Your task in this module: classify snippets from your line of work (emails, staff and customer data, contracts, internal notes, unpublished figures) as GREEN, AMBER or RED, then turn an email into a version you could safely give to an AI tool. The exercise has a variant for each role; the finance variant has 30 snippets, including register extracts and unpublished NAVs. By the end you will make the "can I paste this?" decision in seconds, and you will know what to do when your firm has no written rule.
All data in the exercise is fictional. You may paste it into any tool.
What you will be able to do
- Run the three checks (tool, data, policy) before every prompt
- Describe what Microsoft 365 Copilot, ChatGPT Enterprise and personal accounts do with prompts, history and web search, from the vendors' own documentation
- Explain how professional secrecy (Article 41 of the Law of 5 April 1993), the GDPR, the AML tipping-off rule and the MAR definition of inside information limit what can be pasted
- Classify work content as GREEN, AMBER or RED by its most sensitive item
- Minimise an email with placeholders so an AI tool can draft a reply without identifying data
- Act correctly when there is no written AI policy or after a mistaken paste
5 video lessons, 20 minutes
- 2.1 Three checks before every prompt, and what the tools do with your paste 4:14
- 2.2 The legal frame in five rules 4:20
- 2.3 GREEN, AMBER, RED: the course policy and the never-paste list 4:23
- 2.4 How to get help without pasting the data, and what to do after a mistake 3:54
- 2.5 What is RED in your role, and the checklist before the exercise 3:38
Exercise: Classify what you would paste, then minimise it
A variant for each role, opening on yours: General (any role). Fictional data, scored automatically against the answer key, with the model answer after you submit.
Module quiz
8 questions, pass mark 6. Passing it counts towards your course certificate.
Sources
- Law of 5 April 1993 on the financial sector, Article 41(1) and (2a), CSSF English consolidated text: https://www.cssf.lu/wp-content/uploads/L_050493_lfs.pdf
- Law of 12 November 2004 on AML/CFT, Article 5(5), CSSF English consolidated text: https://www.cssf.lu/wp-content/uploads/L_121104_AML.pdf
- GDPR, Regulation (EU) 2016/679, recital 26, Articles 4(1), 5(1)(c), 9(1), 33(1): https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng
- Market Abuse Regulation, Regulation (EU) No 596/2014, Article 7(1)(a): https://eur-lex.europa.eu/eli/reg/2014/596/oj/eng
- CSSF thematic review on AI, May 2025, page 5: https://www.cssf.lu/wp-content/uploads/Thematic_report_AI_2024.pdf
- CSSF annual report 2025 (French), PDF page 18: https://www.cssf.lu/wp-content/uploads/CSSF_RA_2025_FR.pdf
- Investment Officer, ABBL survey, 11 May 2023: https://www.investmentofficer.lu/en/news/luxembourg-banks-seek-legal-clarity-use-generative-ai
- Microsoft Learn, Enterprise data protection in Microsoft 365 Copilot: https://learn.microsoft.com/en-us/copilot/microsoft-365/enterprise-data-protection
- Microsoft Learn, Data, privacy, and security for Microsoft 365 Copilot: https://learn.microsoft.com/en-us/copilot/microsoft-365/microsoft-365-copilot-privacy
- OpenAI, Enterprise privacy (updated 8 January 2026), Wayback capture of 2 June 2026: https://web.archive.org/web/20260602132611id_/https://openai.com/enterprise-privacy/