Module 45Stage: Approve90 min

Responsible AI

Understand the basic privacy considerations of using AI tools with customer data.

Jump to lesson:
01Privacy02Copyright03Hallucinations04Bias05Brand Safety
1
Lesson 1

Privacy

Objective

Understand the basic privacy considerations of using AI tools with customer data.

Teach
Be deliberate about what customer data you input into AI tools — avoid pasting sensitive personal information into general-purpose AI tools without checking their data handling policies, and understand consent requirements (see the Nigeria NDPA note in Academy 07).
Real-World Example
Summarizing anonymized customer feedback patterns is generally lower-risk than pasting a customer's full personal details into an AI tool.
Common Mistake
Pasting sensitive customer or business data into AI tools without checking their privacy policy.
Practical Exercise

Review what kinds of data you currently input into AI tools, and flag anything sensitive.

Lesson Checkpoint

A safe default when using AI tools with customer data is to

2
Lesson 2

Copyright

Objective

Understand basic copyright risk when using AI-generated content.

Teach
AI-generated text and images can sometimes closely resemble existing copyrighted work — apply the same originality checks you'd apply to any content, and never claim exact wording/style is definitely original without checking.
Real-World Example
Running AI-generated copy through a quick originality check before publishing, especially for anything close to existing well-known phrasing.
Common Mistake
Assuming AI-generated content is automatically free of any copyright concern.
Practical Exercise

Note one step you could add to check AI-generated content for originality before publishing.

Lesson Checkpoint

AI-generated content should be treated regarding copyright as

3
Lesson 3

Hallucinations

Objective

Recognize and catch AI hallucinations (confidently stated false information).

Teach
AI can state incorrect facts, statistics, or quotes with the same confident tone as correct ones — there's no reliable "tell" other than verification, which is why Module 41's human-review habit matters most here.
Real-World Example
An AI confidently citing a statistic that doesn't actually exist anywhere, stated with the same tone as a verified fact.
Common Mistake
Assuming confident-sounding AI output is more likely to be accurate.
Practical Exercise

Find one specific, checkable claim in a recent AI output and verify it.

Lesson Checkpoint

AI hallucinations are dangerous mainly because they

4
Lesson 4

Bias

Objective

Recognize that AI can reflect biases present in its training data.

Teach
AI outputs can unintentionally reflect stereotypes or skewed assumptions present in training data — review AI content (especially about people, cultures or sensitive topics) with this in mind.
Real-World Example
Reviewing an AI-generated customer persona to check it isn't relying on lazy stereotypes instead of real research (Module 44, Lesson 3).
Common Mistake
Assuming AI output is neutral by default and skipping a bias check on sensitive content.
Practical Exercise

Review one piece of AI-generated content involving people or culture for potential bias.

Lesson Checkpoint

AI bias in outputs mainly comes from

5
Lesson 5

Brand Safety

Objective

Set clear boundaries for what AI should never do unsupervised for your brand.

Teach
Define topics or actions that always require human sign-off — sensitive news events, competitor call-outs, pricing claims, legal/medical/financial claims — and encode them as standing rules (like KoraSpace's AI Memory, Section 7.2).
Real-World Example
"Never comment on competitors by name" or "never publish pricing without human confirmation" as standing AI Memory rules.
Common Mistake
Having no defined boundaries, discovering brand-safety issues only after something has already been published.
Practical Exercise

Write 3 brand-safety rules you'd set as standing AI instructions.

Lesson Checkpoint

Brand safety boundaries for AI should be

Module Template

**AI Marketing Governance Checklist** — see Templates Library.

Deepen Your Knowledge
  • •Anthropic's Trust & Safety and usage policy documentation
  • •Nigeria Data Protection Act (NDPA) 2023 overview — see Resource Library
Module Checkpoint Assessment

Test Your Mastery of Module 45

5 questions · 70% passing threshold required

PreviousMod 44: AI Research & Analytics
Next ModuleMod 46: Creator Economy