Pathway 1 · Lesson 0530 min

AI Diligence & Responsible Use

Ethics, governance, and taking ownership of AI output

Learning Objective

By the end of this lesson, you will be able to: Apply the Green/Amber/Red data classification framework, describe three key privacy principles for AI use, and draft a personal AI governance policy.

60%

of organisations that experienced AI incidents report unsatisfactory response

Engage

"AI governance isn't a compliance checkbox. It's the difference between AI that builds trust and AI that destroys it. What does responsible use actually look like in practice?"

Explore

A practical introduction to AI ethics — moving beyond abstract principles to concrete decisions you make every day.

Explain

Data Classification: The Green/Amber/Red Framework

Before inputting anything into an AI tool, classify the data:

🟢 Green — Safe to use: Public information, published research, general knowledge, your own non-sensitive work product. Can be used with any approved AI tool.

🟡 Amber — Use with caution: Internal business information, non-public strategies, aggregated data without personal identifiers. Use only with approved, enterprise-grade AI tools with data protection agreements.

🔴 Red — Never input: Personal data (names, emails, health, financial), confidential client information, trade secrets, legally privileged material. Never input into any external AI tool.

⚠️

The Data Leak Risk Is Real

In 2023, Samsung engineers accidentally leaked proprietary source code and meeting notes by pasting them into ChatGPT. The data was potentially used to train future models. This happened because employees didn't understand data classification. The Green/Amber/Red framework prevents this.

Explain

Privacy Principles for AI Use

Three principles protect both individuals and your organisation:

1. Data Minimisation

Only input the minimum data necessary for the task. If you need AI to draft a performance review, describe the situation in general terms — don't paste in the employee's personal file.

2. Anonymisation

Before inputting data about real people, remove or replace identifying information. 'Employee X in the London office' instead of a person's name.

3. Consent Awareness

Be aware of whether the people whose data you're discussing have consented to AI processing. Client data, customer data, and employee data all carry different consent obligations.

Explain

Disclosure: When Must You Tell Others?

Disclosure requirements vary by context, but a practical framework:

Always disclose when:

The output will be presented as your original work in a professional or academic context

Client contracts or professional standards require it

The output involves significant decisions affecting people's lives

You're asked directly whether AI was used

Consider disclosing when:

AI contributed substantially to the final output

The audience might reasonably want to know

Disclosure doesn't mean apology. Using AI professionally is a skill. Disclose it the same way you'd disclose using any other professional tool.

Personal AI Governance Policy Template

Element	Your Policy
Approved Tools	Which AI tools can I use for work?
Data Rules	What data is Green / Amber / Red for me?
Disclosure	When will I disclose AI use?
Quality Gate	What review do I do before sharing AI output?
Prohibited Uses	What will I never use AI for?
Escalation	Who do I contact if something goes wrong?

Try It · Reflect & Apply

Draft your personal AI governance policy using the template above. Be specific — generic policies don't change behaviour. What are your actual approved tools? What data do you regularly work with that needs classification?

Your notes are saved in your browser for this session.

Evaluate · Knowledge Check

1 / 3

A colleague wants to paste a client's financial projections into ChatGPT to generate a summary. How should you classify this data?

Fluency Practice Studio

Responsible Use Studio

Make the decision

Decide whether the information in front of you is Green, Amber, or Red before entering it into any AI system.

Practise at work

Classify five real items from your work: a public webpage, a team meeting summary, a client brief, an employee performance note, and a spreadsheet with contact details. Write the safe handling action for each.

Quality check

Can you state which approved tool, disclosure action, and review step apply to the information? If you are unsure, stop and seek the organisation’s policy owner.

Spaced revisit

For the next seven days, keep a private tally of one data-classification decision each day. Review the patterns with a trusted colleague or champion.

External video references

Keep Learning on YouTube

Open in a new tab

AI Ethics and Responsible UseEmbedded lesson videoA practical introduction to AI ethics — moving beyond abstract principles to concrete decisions you make every day.Watch · Watch on YouTube

Ready to move on?

Mark this lesson complete and continue to "Building AI Culture in Your Team".