How exposed is your business to AI risk?
24 plain-English questions in 8 areas, each linked to the ISO/IEC 42001 clause or control it relates to. See your score, the gaps that matter most and what to fix first. It takes about 8 minutes.
1.AI inventory and visibility
Do you have a list of every AI tool used in the business, including AI features built into software such as Microsoft 365, your CRM or helpdesk?
ISO/IEC 42001: Clause 4.3 · A.4.2
For each AI tool, do you know whether you build it, buy it, or simply use it?
ISO/IEC 42001: Clause 4.3 · A.10.2
Do you know which business decisions or customer interactions depend on AI output?
ISO/IEC 42001: A.5.2 · A.9.4
2.Policy and accountability
Do you have a written AI policy approved by senior leadership?
ISO/IEC 42001: Clause 5.2 · A.2.2
Is a named person accountable for AI risk in the organisation?
ISO/IEC 42001: Clause 5.3 · A.3.2
Does leadership review AI use and AI risks at least once a year?
ISO/IEC 42001: Clause 9.3 · A.2.4
3.Staff use and shadow AI
Have staff been told which AI tools they may use, and which they must not?
ISO/IEC 42001: A.9.2 · A.9.4
Are there clear rules on what must never be entered into AI tools, such as client data, personal data, source code or financials?
ISO/IEC 42001: A.9.2 · A.7.2
Have staff been trained on AI risks, such as made-up answers and data leakage?
ISO/IEC 42001: Clauses 7.2 and 7.3 · A.4.6
4.Data and privacy
Do you know whether your AI suppliers store your prompts or use them to train their models?
ISO/IEC 42001: A.7.5 · A.10.3
Have you carried out a data protection impact assessment (DPIA) for AI that processes personal data?
ISO/IEC 42001: A.5.4 · UK GDPR Article 35
Do you know which countries your AI tools process your data in?
ISO/IEC 42001: A.7.5 · A.10.3
5.Customer-facing AI and accuracy
If customers interact with a chatbot or AI tool, are they clearly told they are dealing with AI?
ISO/IEC 42001: A.8.2 · A.8.5
Is AI output checked by a person before it reaches customers or is used to make decisions?
ISO/IEC 42001: A.9.4 · A.6.2.6
Do you regularly test your AI for wrong, biased or inappropriate answers?
ISO/IEC 42001: A.6.2.4 · A.5.4
6.Security and misuse
Have you tested whether your AI tools can be tricked into ignoring instructions or revealing data (prompt injection)?
ISO/IEC 42001: A.6.2.4
Are AI conversations and actions logged so you could investigate if something went wrong?
ISO/IEC 42001: A.6.2.8
Is AI use controlled through company-managed accounts rather than personal free accounts?
ISO/IEC 42001: A.4.4 · A.9.2
7.AI suppliers
Did you assess AI suppliers' security, privacy and AI practices before you started using them?
ISO/IEC 42001: A.10.3
Do your contracts say who is responsible if AI output is wrong or causes harm?
ISO/IEC 42001: A.10.2
Would you know if a supplier changed the AI model or its terms of use?
ISO/IEC 42001: A.10.3 · A.6.2.6
8.Incidents and risk management
Can staff and customers easily report an AI mistake or concern?
ISO/IEC 42001: A.3.3
Do you have a plan for responding to an AI incident, such as a chatbot giving harmful advice in public?
ISO/IEC 42001: A.8.4 · Clause 10.2
Have you carried out a risk assessment that specifically covers AI?
ISO/IEC 42001: Clauses 6.1.2 and 6.1.4
This self-assessment is a guide to ISO/IEC 42001 and common AI risks, not an audit, legal advice or certification. Certification is carried out by an independent accredited certification body, not BK17.