Does ISO/IEC 27001:2022 allow or prohibit AI coding tools?

ISO/IEC 27001:2022 defines requirements for an information security management system; the public ISO description does not decide a named AI coding tool. An organization determines scope, assesses risk, selects and implements controls, evaluates effectiveness, and retains evidence. Applicability and conformity remain organization- and assessor-specific.

Does an AI coding policy prove ISO/IEC 27001 conformity?

No. A policy states an approved rule. A technical or organizational control implements part of that rule, and an evidence artifact records an observed event. Those are inputs to an ISMS and an audit, not a certification decision by themselves.

Does a verification report count as audit evidence?

It can be one evidence artifact if it is authentic, scoped, retained, and relevant to a selected control. The auditor still evaluates the defined ISMS scope, control design, implementation, operation, and other evidence. A report does not certify the organization or the tool.

Is TISAX the same as ISO/IEC 27001 certification?

No. ENX describes TISAX as an assessment and exchange mechanism based on the VDA ISA, with assessment objectives defining applicable requirements. ISO/IEC 27001 is an ISMS requirements standard; certification is a separate conformity-assessment route. Check the current TISAX objective, ISA version, scope, and partner request.

What should an AI-coding control owner document?

Document the risk and decision first: approved tools and configurations, permitted data classes, identities and access, supplier and model endpoints, required checks, exceptions, owners, retention, and review cadence. Then retain evidence that the selected controls operated. This is an implementation pattern, not a prescribed ISO control set.

Compliance

ISO/IEC 27001 and AI Coding

Last updated: 2026-07-173 min read

ISO/IEC 27001:2022 governs an ISMS, not a named AI-tool list. For AI coding, the defensible work is to define scope, assess risk, select and implement controls, and retain relevant evidence. A policy, technical check, or evidence report may support that work; none is a certification or audit result.

Contents

Start with the standard's actual boundary

ISO’s current public record identifies ISO/IEC 27001:2022 as the third-edition ISMS requirements standard, published in October 2022, with Amendment 1:2024. ISO also says organizations may implement the standard without seeking certification. Certification is a separate external conformity-assessment route. Checked 17 July 2026; the full licensed standard, the organization’s scope, and the assessor’s criteria remain authoritative.

Therefore this article does not claim that ISO/IEC 27001 or TISAX categorically permits or forbids AI coding. It maps AI-tool risk questions to management-system work. Your ISMS owner selects the applicable controls; your certification body or TISAX audit provider evaluates the agreed scope.

Source and applicability boundary

Scope: ISO/IEC 27001:2022 and Amendment 1:2024, ISO's public certification boundary, BSI/ANSSI guidance, and ENX's TISAX process description.
As of: 17 July 2026
Method: Separate what a primary source states from an editorial implementation example, and keep policy, control, evidence, audit, and certification as distinct layers.
Primary sources: ISMS requirements: ISO describes ISO/IEC 27001:2022 as requirements for establishing, implementing, maintaining, and continually improving an ISMS.ISO/IEC 27001:2022 · 2026-07-17The full normative standard is licensed; this article relies on ISO's public description and does not reproduce or interpret clauses as audit advice.
Certification boundary: ISO states that organizations may implement the standard without certification and that certification is performed through an external certification body.ISO management-system standards · 2026-07-17A certificate's scope, issuing body, accreditation, and current status must be checked separately.
TISAX boundary: ENX describes assessment objectives, applicable ISA requirements, an audit-provider assessment, and a resulting TISAX label.ENX TISAX Participant Handbook · 2026-07-17TISAX is not treated here as an ISO/IEC 27001 certificate or as a blanket approval of AI tools.
Not inferred: No certification or audit-readiness conclusion
No universal control selection for AI coding
No claim that local or cloud processing is inherently conforming
Limitations: The organization defines its ISMS scope and risk treatment
Applicable contractual, TISAX, customer, sector, and legal requirements vary
An auditor or assessment provider evaluates conformity within the agreed scope

The cited sources define management-system and assessment structures; they do not approve a particular AI tool or certify a reader's organization.

A practical implementation example

For one repository, define which assistants and model endpoints are approved, which data classifications may reach them, how identities and access are managed, which checks are required before merge, who can approve exceptions, and how the decision is reviewed. The 2024 BSI/ANSSI guidance on AI coding assistants is a useful risk input, but it is guidance-not a certification decision or a prescribed ISO control set.

Do not collapse the assurance layers

A useful artifact can support the next layer without proving it; certification remains an external, scoped decision.
	What it is	AI-coding example	What it does not prove
ISMS requirement	ISMS requirementLayer distinction · 2026-07-17Plain-language orientation, not normative terminology.	A requirement the organization addresses in its scoped management systemImplementation exampleThe organization selects controls from its own risk treatment and requirements.	That a particular tool or control is mandatoryClaim boundary · 2026-07-17External assessors and certification bodies make their own scoped determinations.
Internal policy	Internal policyLayer distinction · 2026-07-17Plain-language orientation, not normative terminology.	Approved rules for tools, data classes, accounts, and checksImplementation exampleThe organization selects controls from its own risk treatment and requirements.	That the rule is implemented or effectiveClaim boundary · 2026-07-17External assessors and certification bodies make their own scoped determinations.
Implemented control	Implemented controlLayer distinction · 2026-07-17Plain-language orientation, not normative terminology.	Managed identity, data-path restriction, review gate, or exception processImplementation exampleThe organization selects controls from its own risk treatment and requirements.	That every risk is treated or every run followed itClaim boundary · 2026-07-17External assessors and certification bodies make their own scoped determinations.
Evidence artifact	Evidence artifactLayer distinction · 2026-07-17Plain-language orientation, not normative terminology.	Configuration record, approval, log, test result, or verification reportImplementation exampleThe organization selects controls from its own risk treatment and requirements.	That the ISMS conforms or an audit will passClaim boundary · 2026-07-17External assessors and certification bodies make their own scoped determinations.
External audit work	External audit workLayer distinction · 2026-07-17Plain-language orientation, not normative terminology.	Sampling and evaluation against the defined scope and criteriaImplementation exampleThe organization selects controls from its own risk treatment and requirements.	A certification result before the auditor concludesClaim boundary · 2026-07-17External assessors and certification bodies make their own scoped determinations.
Certification decision	Certification decisionLayer distinction · 2026-07-17Plain-language orientation, not normative terminology.	A scoped decision by the certification bodyImplementation exampleThe organization selects controls from its own risk treatment and requirements.	Security of every product, change, supplier, or future operationClaim boundary · 2026-07-17External assessors and certification bodies make their own scoped determinations.

A useful artifact can support the next layer without proving it; certification remains an external, scoped decision.

TISAX is a separate assessment path

The ENX Participant Handbook describes TISAX assessment objectives, the applicable VDA ISA requirements, self-assessment, audit-provider work, and resulting labels. A customer’s requested objective, the current ISA version, locations, protection needs, and assessment scope determine what is evaluated. An ISO/IEC 27001 certificate, a TISAX label, and an internal AI-tool policy are related evidence contexts, not interchangeable outcomes.

Where Reality Graph fits

Reality Graph can retain one bounded chain from written task to selected checks and observed results. How that chain is computed into one outcome, and what is written down when part of it is missing, is documented on the product side. That may support a control owner or auditor when it is relevant, authentic, and inside the defined scope. It does not select the organization’s controls, operate the whole ISMS, certify conformity, or replace external assessment. See the broader distinction in audit trails for AI code.

This workflow can record

The approved task, scope, and selected checks
Observed command results and changed files for one governed run
The operator decision and preserved evidence boundary

It cannot establish

Organization-wide conformity with ISO/IEC 27001:2022
A TISAX assessment result or label
Certification, audit success, control effectiveness beyond the evidence, or legal compliance

Reality Graph can produce bounded implementation evidence; it cannot make an ISO/IEC 27001 or TISAX determination.

FAQ

Does ISO/IEC 27001:2022 allow or prohibit AI coding tools?: ISO/IEC 27001:2022 defines requirements for an information security management system; the public ISO description does not decide a named AI coding tool. An organization determines scope, assesses risk, selects and implements controls, evaluates effectiveness, and retains evidence. Applicability and conformity remain organization- and assessor-specific.
Does an AI coding policy prove ISO/IEC 27001 conformity?: No. A policy states an approved rule. A technical or organizational control implements part of that rule, and an evidence artifact records an observed event. Those are inputs to an ISMS and an audit, not a certification decision by themselves.
Does a verification report count as audit evidence?: It can be one evidence artifact if it is authentic, scoped, retained, and relevant to a selected control. The auditor still evaluates the defined ISMS scope, control design, implementation, operation, and other evidence. A report does not certify the organization or the tool.
Is TISAX the same as ISO/IEC 27001 certification?: No. ENX describes TISAX as an assessment and exchange mechanism based on the VDA ISA, with assessment objectives defining applicable requirements. ISO/IEC 27001 is an ISMS requirements standard; certification is a separate conformity-assessment route. Check the current TISAX objective, ISA version, scope, and partner request.
What should an AI-coding control owner document?: Document the risk and decision first: approved tools and configurations, permitted data classes, identities and access, supplier and model endpoints, required checks, exceptions, owners, retention, and review cadence. Then retain evidence that the selected controls operated. This is an implementation pattern, not a prescribed ISO control set.

Keep reading

ComplianceAI Coding in Regulated IndustriesDORA, IEC 62304/MDR, ISO 26262/ASPICE regulate process and evidence, not code authorship. What each sector demands when AI writes code - and the one base setup that fits all three.ComplianceAudit Trails for AI-Generated CodeGit records what changed - not what the task was, what was validated, who approved. The five auditor questions, why the demand rises under NIS2 and the new product liability, and how to build the trail.All articlesThe whole collection - 58 cited, dated guides on verifying AI-generated code.

Sources

Want to see what your last agent run would have looked like?

Request access