Skip to content
arrow_back
search
Annex A 9.2 psychology ISO/IEC 42001:2023

Processes for Responsible Use of AI Systems

Set clear objectives to ensure responsible use of AI systems.

record_voice_over

Plain language

This control is about setting goals for using AI responsibly so it works in ways that help and don’t harm. Imagine if your AI accidentally sent a wrong bill to a customer or showed them a product that doesn't exist - setting the right objectives can prevent those issues.

Framework

ISO/IEC 42001:2023

Control effect

Preventative

Classifications

N/A

Official last update

01 Dec 2023

Control Stack last updated

19 May 2026

Maturity levels

N/A

Official control statement

The organisation shall define and document the processes for the responsible use of AI systems.
psychology ISO/IEC 42001:2023 Annex A 9.2
priority_high

Why it matters

Without clear objectives, your AI could generate mistakes such as recommending products to customers that don't suit their needs or even offend them, hurting your reputation.

settings

Operational notes

Make it a habit to check if the AI is hitting the responsible goals you set before adopting any updates or changes.

build

Implementation tips

  • The AI lead should start by outlining specific goals for each AI system, such as reducing customer complaints or improving delivery accuracy. They can have a simple list or spreadsheet to track if these goals are being met.
  • Board members should ensure there's a company policy that spells out ethical AI use. A short clause stating the intention to prevent harmful outcomes can guide everyone involved.
  • Procurement should ask AI vendors how their solutions align with your goals. Adding a requirement in supplier contracts that their AI services won’t violate your objectives ensures alignment from the start.
  • The head of risk should regularly review whether the AI is meeting its intended objectives. They can use a quarterly meeting to look at reports showing how AI impacts the business, such as customer satisfaction metrics.
  • The data steward should verify that training data supports responsible goals. Check that the data don’t have biases by sampling data sets or running a diversity analysis, which ensures a fair AI model output.
fact_check

Audit / evidence tips

  • AskRequest the list of AI objectives set by the organisation. GoodThe organisation has a clear list of objectives geared towards responsible AI use.
  • AskAsk for the vendor contract documents. GoodContracts have clauses ensuring vendor AI practices align with company objectives.
  • AskInquire about the AI policy document. GoodThe policy clearly outlines ethical AI use and objectives based on responsible AI practices.
  • AskReview meeting minutes or reports from risk assessments. GoodRegular reviews show evaluations against set AI objectives and document outcomes.
  • AskRequest to see training data assessments. GoodThe data used is checked for biases, with outcomes recorded to ensure responsible AI delivery.
link

Cross-framework mappings

How Annex A 9.2 relates to controls across ISO/IEC 27001, ISO/IEC 42001, Essential Eight, and ASD ISM.

ISO 27001

Control Notes Details
handshake Supports (1) expand_less
Annex A 5.4 Annex A 5.4 requires management to ensure personnel apply established policies, supporting Annex A 9.2's AI-use processes
extension Depends on (1) expand_less
Annex A 5.1 Annex A 9.2 mandates defining and documenting processes for responsible AI use

ASD ISM

Control Notes Details
sync_alt Partially overlaps (1) expand_less
ISM-2074 Annex A 9.2 requires documenting processes for responsible AI use, broader than ISM-2074's policy focus
handshake Supports (1) expand_less
ISM-1999 ISM-1999 aligns cyber security strategy with business goals, supporting Annex A 9.2 by ensuring AI-use processes reflect organisational o...

These mappings show relationships between controls across frameworks. They do not imply full equivalence or certification.

psychology

Want to implement this AI control?

Mindset Cyber runs PECB-accredited ISO/IEC 42001 training that maps directly to the AI controls in this library.

Mapping detail

Mapping

Direction

Controls