Domain 6 – Governance, Risk, and Responsible Use (15%)

• Identify appropriate vs inappropriate use cases
• Apply data sensitivity, privacy, and regulatory considerations
• Follow organisational AI policies
• Understand ethical implications of AI use in the organisation

CCAO-F Domain 6 · Task 6.1

What You Need to Know

Two questions decide whether a task goes to Claude. Fit (Domain 4) asks whether Claude would do the work well. Appropriateness asks whether the task should be handed over at all.

The Four Inappropriate Categories

1. Work that needs professional authority — legal, medical, financial and safeguarding judgements. Claude can help prepare, summarise or understand; it cannot be the source of the advice. Usage Policy High-risk Areas require a qualified person to review before advice is delivered or a decision is made.
2. A consequential decision about identifiable people — who gets the job, the award, the sanction, the visit. Nobody is accountable for the decision, and the person affected cannot challenge it.
3. Fabrication presented as real — invented quotes, testimonials, case studies or evidence.
4. Anything the Usage Policy or the organisation’s own policy prohibits.

The Substitution Test

Picture a capable, well-meaning colleague who has no professional qualification in the area, no access to the person’s history, and no accountability for the outcome. Would you accept this piece of work from them?

Most Work Is Appropriate

Drafting, reformatting, analysing material you hold, and preparing someone to make a decision are all appropriate. The inappropriate categories are narrow and specific. Treating everything near them as forbidden abandons work a simple safeguard would enable.

Key Concept: Adding review to an inappropriate delegation does not repair it. Review cannot convert a decision that should not have been delegated into one that should.

Exam Traps

• Adding a human reviewer to an inappropriate delegation and treating that as the fix
• Accepting the internal-use excuse
• Declining both halves of a mixed proposal

Practice Scenario

Children’s centre manager is asked to use Claude on the annual family survey. Two proposals: summarise 300 free-text responses into themes for the service review, and score individual families’ responses to flag which should receive a home visit. Centre policy allows Claude on the survey text and bars any tool from scoring families for home visits. What should the manager do?

Build Exercise: Sort a Fortnight of Real Work (25 min)

CCAO-F Domain 6 · Task 6.2

What You Need to Know

Sensitive material is anything an organisation should not hand to an outside service without a decision. A safeguard changes what leaves the organisation’s hands.

Recognising Sensitive Material

Names, addresses, dates of birth, reference and account numbers, health information, financial details, anything about children — plus categories about ownership and promises: duty of confidence, commercially confidential figures, third-party material governed by a contract.

The Ladder (Lowest First)

1. Remove — strip the identifiers and proceed
2. Anonymise or aggregate — cohorts instead of individuals, ranges instead of exact figures
3. Use an authorised route — sanctioned workspace under an agreement that covers this class of material
4. Seek authorisation — take the question to whoever owns the data or the policy (an action, not a refusal)

Abandonment sits at the bottom and is almost never keyed.

Why the Placebo Fails

The policy control is about what leaves the organisation’s hands, which happens at upload. Every placebo fails this way: telling Claude to forget, declaring the upload confidential, adding a confidentiality line, deleting the conversation. Nothing typed into the chat can retroactively govern what was uploaded.

Key Concept: Any option whose safeguard operates after the upload is wrong. No placebo is ever keyed. The keyed answer changes what goes in, not what happens afterwards.

Exam Traps

• Treating a confidentiality instruction or later deletion as the safeguard
• Escalating up the ladder when removal would do
• Clearing a document because it contains no names
• Dropping the analysis because personal data is involved

Practice Scenario

Minibus driver training scheme coordinator wants to use Claude to understand why pass rates dropped. Trainee records contain names, licence numbers, dates of birth and medical fitness declarations. Scheme’s data policy allows personal or health details into an AI tool only where the task cannot be done without them. Most appropriate way to proceed?

Build Exercise: Walk the Ladder on One Real Dataset (30 min)

CCAO-F Domain 6 · Task 6.3

What You Need to Know

The rule is short: use the workspace your organisation provides.

Why the Sanctioned Workspace Is Substance

When an organisation provides Claude, the arrangement governs how its data is handled. The customer is Controller; Anthropic is Processor on the customer’s behalf. By default Anthropic will not use inputs or outputs from commercial products to train its models. Agreement, administration and data handling all attach to the sanctioned workspace.

Personal Accounts Are Different

A personal account sits under the individual’s own consumer relationship with Anthropic. The employer does not administer, see or govern any of it. A paid personal plan with careful settings changes nothing — the question is whose agreement governs the data.

The Shadow-AI Response Has Two Halves

1. Do not copy the workaround, however pressing the deadline
2. Raise it with whoever owns the policy or the data

Deleting chats or closing the account is after-the-fact tidying. The disclosure has already happened.

Policy Silence and Working Habits

A policy’s silence means the question has not been answered. Ask the policy’s owner. When policy conflicts with convenience, the policy wins. When the policy is genuinely wrong or too slow, get its owner to change it rather than work around it.

Key Concept: The keyed answer does both halves of the shadow-AI response. Candidates reliably miss the second.

Exam Traps

• Fixing only half of the shadow-AI situation
• Offering deletion as remediation for material already disclosed
• Treating policy silence as permission

Practice Scenario

Operations lead at a recycling social enterprise discovers a colleague has been pasting the local authority contract and staff shift records into a personal free Claude account at home, because the organisation’s licences have not been issued yet. Enterprise’s data rules keep its material out of personal accounts on any plan, and any slip must be reported. What should the operations lead do?

Build Exercise: Map Your Own Governance Position (25 min)

CCAO-F Domain 6 · Task 6.4

What You Need to Know

This objective asks who owns the output, who deserved to know how it was made, and who gets described by it.

Accountability

The person whose name is on the work owns it. Saying Claude drafted it describes the production method and moves no responsibility. In Usage Policy High-risk Areas, a qualified person must remain accountable for what is delivered or decided. (AI Fluency: Diligence.)

When Disclosure Is Owed

Categorical rule (Usage Policy): Consumer-facing chatbots and external-facing AI agents must tell users they are interacting with AI rather than a human — at a minimum at the beginning of each chat session or in the product interface.

Everything else runs on proportion. Disclosure is owed when:

• The reader’s trust depends on who did the thinking or the judging (assessed work, professional opinion)
• A claim of authorship would be false
• A policy or client agreement requires it

A personal message such as condolence is different — its reader relies on you having done the feeling, so you write it yourself.

Fairness Is a Different Review from Accuracy

An output can pass every accuracy check and still be unfair. Descriptions of groups can repeat assumptions the facts never required. It needs a different reviewer: someone who knows those people, reading for how they are described.

Colleagues Are Stakeholders Too

When Claude takes over part of a process, the colleagues doing that work need to know what changed and why. Using Claude to produce a colleague’s contribution without telling them is an ethical question of honesty between colleagues.

Key Concept: A 6.4 key is always about ownership, disclosure, fairness or colleagues — never about checking or validating (Domain 2).

Exam Traps

• Transferring accountability to the tool or its configurer
• Using disclosure as a shield or as a blanket rule
• Signing off work about people once the figures check out

Practice Scenario

Clerk of an almshouse charity publishes the annual report to residents and donors, largely drafted by Claude from the charity’s own records. A resident spots that the report overstates how quickly repairs were completed last year, and asks the clerk who wrote it. What is the professional response?

Build Exercise: Audit One Published Piece (30 min)