Your responsible AI habits
Responsible AI use isn't a rulebook — it's a set of habits. This module turns everything you've learned into a routine you can use every day.
Learning objectives
- Explain why accountability stays with people
- Describe the EU AI Act's AI literacy expectation
- Decide when to disclose AI use
- Describe what makes human oversight meaningful
- Apply the Decide–Protect–Describe–Check–Own habit loop
6.1 You are accountable
AI can draft, suggest and even act — but responsibility stays with people. If you send an AI-written email, publish an AI-generated report or act on an AI recommendation, it's your work. "The AI said so" is not a defence to a customer, a regulator or a court.
6.2 AI literacy is now an expectation
Under Article 4 of the EU AI Act, organisations that provide or use AI systems must take measures to support a sufficient level of AI literacy among their staff, taking into account people's knowledge, experience and the context in which AI is used. This has applied since 2 February 2025, and it affects UK organisations that offer or use AI in the EU market. Even where the law doesn't require it, employers increasingly expect it — in 2026 UK employers named AI skills as the hardest to find for the first time (ManpowerGroup).
The EU AI Act Fundamentals course explains the wider law.
6.3 Transparency: when to say you used AI
There's no single rule for every situation, but these principles work well:
- Follow your organisation's policy and any client, publisher, academic or regulator rules.
- Disclose when it would matter to the reader — for example, if they'd assume a person wrote or checked something, or if AI shaped a decision about them.
- Never pass off AI-generated images, audio or video of real people as genuine. Under the EU AI Act, deepfakes must be disclosed.
- Tell people when they're talking to an AI if you deploy a chatbot to customers.
6.4 Human oversight that actually works
"Human in the loop" only helps if the human can genuinely question the AI. Good oversight means:
- The reviewer understands what the AI does and where it tends to fail
- They have the time and authority to override it
- Higher-stakes decisions get more scrutiny — and some decisions shouldn't be delegated to AI at all
- Mistakes are reported and fed back, so the process improves
6.5 Your personal AI habits
Pulling the course together, here's a habit loop you can use every day:
- Decide — Is this a good task for AI? Can I check the result? (Module 3)
- Protect — Approved tool, minimal data, no secrets. (Module 4)
- Describe — Task, context, format and constraints. Iterate. (Module 3)
- Check — Run CHECK on the output before using it. (Module 5)
- Own — Edit it into your work, disclose where appropriate, and stay accountable. (This module)
6.6 Keep learning
AI tools change monthly. Set aside a little time to try new features on low-risk tasks, share what works with colleagues, and keep your organisation's guidance in view. The rest of the AI Foundations path builds directly on this course: how AI works, prompt engineering, and validating AI output.
Write five rules you'll follow — for example: "I never paste customer details into AI", "I verify every number before it leaves my desk", "I tell clients when AI drafted a deliverable". Share them with your team and compare.
Key takeaways
- You are accountable for AI-assisted work you use or share.
- Supporting staff AI literacy has been an EU AI Act expectation since February 2025.
- Disclose AI use when it would matter to the reader, and always for deepfakes and customer chatbots.
- Oversight only works if reviewers understand the AI and have time and authority to override it.
- Decide, Protect, Describe, Check, Own.