Home Solutions Insights Case Studies About Book Discovery Call
Governance8 min read

AI Safety vs AI Governance — What a Board Is Really Being Asked

Dr. Mahdi Seify
Dr. Mahdi Seify
Founder & CAIO, VisionXY7 Ltd · Published 14 September 2026

PhD, AI-Driven Business Analytics · ISO/IEC 27001 Lead Auditor and Lead Implementer. Written from delivery, not from a summary of the standard.

In short: When someone asks a board whether its AI is safe, they are not asking whether a policy exists. They are asking whether the AI in the business could harm the people it touches — and whether the organisation would find out if it did. Governance is the mechanism for producing that answer. It is not the answer. An organisation can hold a complete set of governance artefacts and still be unsafe, and the gap between the two is where most real incidents live.

Two words that are not synonyms

Governance is structure. Who decides, who approves, who can stop a thing, what is written down, how often it is reviewed. It is observable, auditable, and it produces artefacts — a policy, a register, a committee, a Statement of Applicability.

Safety is an outcome. It is the claim that the AI operating in your organisation does not harm the people it touches, and that if it started to, you would find out. It produces no artefacts of its own. You can only infer it from evidence.

The relationship between them is one-directional and imperfect. Good governance makes safety more likely. It does not deliver it, and the assumption that it does is the single most expensive mistake we see in this work.

How an organisation ends up governed and unsafe

None of what follows is hypothetical or unusual. Each is a pattern that turns up repeatedly, in organisations that would describe themselves as having AI governance in place.

The policy nobody has read

Approved, dated, published on the intranet. Ask five people who use AI daily what it permits and you get five answers, three of them wrong. The artefact exists; the behaviour it was meant to produce does not.

The approval gate everyone routes around

There is a process for approving an AI system. It takes six weeks. A department needed something in two days, so they used a free tool with a personal account, and it now touches customer data. The gate is real. So is the shadow estate beside it.

The inventory that stopped being true

A maintained AI system inventory, last genuinely reconciled eleven months ago. Every vendor in your stack has shipped AI features since. Nobody added them because nobody procured them — they arrived in a release note.

Human oversight in name

A human reviews every output before it takes effect. That human reviews four hundred a day and agrees with the system ninety-eight percent of the time. That is not oversight; it is a rubber stamp with a job title, and an auditor will say so.

No way to reconstruct a decision

Someone was refused something eight months ago. They have complained. The model has been updated twice, the prompt three times, and no record ties the decision to the version that made it. You cannot investigate your own decision.

In every one of those cases the governance artefacts would pass inspection. In every one, the question a board is actually being asked has no good answer.

The six questions

These are the ones worth putting to your own executives. They are deliberately about evidence and behaviour rather than documents, and none of them can be answered by producing a policy.

  1. Which AI systems make or materially influence decisions about people? Customers, patients, applicants, employees. Not a count of tools — a list of the ones where a person is on the receiving end.
  2. For each of those, what happens when it is wrong? Who is affected, how badly, and how quickly would anyone know.
  3. Who can stop one, today, without asking permission? A name. If the answer is a committee, the answer is nobody.
  4. When did a human last disagree with one of these systems and win? If never, the oversight is nominal and you have just found that out cheaply.
  5. Could we reconstruct a specific AI-assisted decision from six months ago? The inputs, the version, the reviewer, the outcome.
  6. What has gone wrong, and what changed because of it? An organisation with no recorded AI incidents is not a careful one. It is one that is not looking.

An executive team that can answer all six with evidence has something worth calling safety. One that answers all six by describing a process has governance, and a gap.

Why this lands on the board specifically

Oversight is a duty that cannot be delegated. A director can delegate the work of AI governance; they cannot delegate the responsibility for being satisfied it is working. That distinction is familiar from financial controls and health and safety, and it applies here in the same way.

The regulatory direction reinforces it. The EU AI Act places obligations on providers and deployers of high-risk systems — risk management, data governance, human oversight, logging, transparency — and applies on the basis of where the output is used, so a UK organisation whose AI output reaches people in the EU can be in scope. Stand-alone high-risk obligations apply from 2 December 2027 and product-embedded ones from 2 August 2028; those dates are fixed rather than conditional on standards being ready. In the UK the approach is regulator-led rather than statutory — ICO, MHRA, FCA, CQC, GDC — which means no single deadline and no single exemption either.

What to do about it

Build the governance. It is necessary, and there is no route to safety that skips it. Then do the part that is usually skipped: test whether the structure produced the outcome. Take the six questions to your executives and ask for evidence, not process descriptions.

And start by finding out where you actually are, which costs nothing and takes a quarter of an hour.

This article is general information, not legal advice. Certification against ISO/IEC 42001 is issued only by a certification body accredited under ISO/IEC 42006. VisionXY7 Ltd prepares organisations for certification audits and reviews their systems independently; it does not perform them.

The service behind this article
AI Safety & Assurance Programme

One accountable structure across both assessments, the ISO/IEC 42001 gap work, impact assessment, policy and controls, board briefing and the first quarter of Fractional CAIO oversight. From £18,000.

See the programme →

Frequently Asked

Is this just a semantic distinction?

It has a practical test. Ask for your AI policy and you are testing governance. Ask what happened the last time an AI-supported decision was wrong, who noticed, how, and what changed as a result — and you are testing safety. Many organisations can answer the first question in ten seconds and cannot answer the second at all. That gap is not semantic.

Does ISO/IEC 42001 certification mean we are safe?

No. It means an independent body found you operating a management system that meets the standard. That is a strong signal and a genuinely useful one — but the standard certifies the mechanism, not the outcome. A certified organisation with a stale system inventory and an approval gate everybody routes around is certified and not safe. The certificate is evidence, not a conclusion.

Who should own AI safety in an organisation?

One named person with authority to stop a system, reporting to the board on a defined cycle. The title matters less than the two properties. The common failure is distribution: risk owns some of it, legal some, data protection some, IT some, and nobody owns the whole. A question with four owners has none.

What about AI safety in the research sense — alignment, frontier models?

Different field, and worth being clear about. The frontier-safety literature concerns the behaviour of very capable models and is largely the province of the labs building them. What a board of an ordinary organisation faces is nearer-term and more mundane: a procurement tool screening out applicants unfairly, a triage assistant missing a case, a chatbot committing the company to something. Both are real. Only the second is on your risk register.

Where should a board start?

With the six questions in this article, put to their own executives, and with one of the free assessments. The assessment takes under fifteen minutes and produces something concrete to discuss. If it says the honest answer is to revisit in six months, that is a legitimate finding and a cheap one.

Would You Know
If It Went Wrong?

The free AI Readiness assessment takes under twelve minutes and scores your organisation on exactly this — whether anyone owns AI, whether a human reviews output before it takes effect, and whether an AI-assisted decision could be reconstructed six months later.

Take the free assessment