Standards & Regulation

No single standard. One coherent map.

There is no global rulebook for AI, and dozens of jurisdictions are writing their own. The instruments that matter overlap as much as they diverge: binding regulation, risk frameworks, certifiable management systems, and the harmonised standards, one cited and the rest still being drafted. They also keep moving. Nobody needs to know all of them. What matters is which bind you, which are still in flight, and how they fit together.

Written against
ISO 10005:2018 · ISO 21502:2020 · ISO 22301:2019 · ISO 31000:2018 · ISO 9001:2026 · ISO/IEC 23894:2023 · ISO/IEC 38500:2024 · ISO/IEC 38507:2022 · ISO/IEC 42001:2023 · ISO/IEC 42005:2025 · ISO/IEC 42006:2025 · NIST AI RMF 1.0 · IEEE 7000-2021 · Regulation (EU) 2024/1689 · EN 18286:2026 · ISO/IEC 27090:2026
Last updated
Answered first

What is each of the standards for, and which apply to us?

Nobody needs to know all of them: regulation sets what you must do, and the standards and frameworks give you ways to do it. ISO/IEC 42001, 42005 and 42006 cover AI management systems, impact assessment and the bodies that certify them. NIST AI RMF, ISO/IEC 23894 and ISO 31000 cover risk. ISO/IEC 38507 covers oversight by the governing body. Which apply depends on your role, your jurisdiction and what your customers ask for.

Which of these are binding and which are voluntary?

The EU AI Act and national laws bind; the ISO standards and the NIST AI RMF are voluntary ways to meet them. The NIST AI RMF is not certifiable. ISO/IEC 42001 is a certifiable management system standard that gives no presumption of conformity under the AI Act. A harmonised standard does give that presumption, but only once it is cited in the Official Journal.

Which standards matter if we are a provider, and which if we are a deployer?

A provider of high-risk AI systems needs the Article 17 quality management evidence; a deployer is not reached by Article 17. For a provider, EN 18286:2026 is the standard written for Article 17 and ISO/IEC 42001 is the certifiable governance base. For a deployer, ISO/IEC 42001 and, where a quality system exists, ISO 9001 are the relevant management frameworks.

The instruments are not interchangeable, and the differences decide how each is used. Regulation sets what you must do, and differs sharply by jurisdiction. Risk frameworks set how to organize the work. Management systems make the result auditable.

Regulation

EU AI Act

The first comprehensive horizontal AI law. Risk-tiered, enforceable, extraterritorial. Binding, and the one with penalties.

How it lands in your management system

US state law

No federal statute, so the map is state by state. Texas TRAIGA reaches developers and deployers generally. California SB 53 reaches frontier-model developers, with its heaviest duties triggered only above a revenue and compute threshold. Both applied from January 2026; Colorado’s replacement act applies from January 2027.

Canada — no AI statute

What binds instead is privacy law and sector supervision: PIPEDA, Quebec’s Law 25, and OSFI Guideline E-23 for federally regulated financial institutions. Covered in full under supervision below.

Harmonised standards for the AI Act

Written by CEN-CENELEC's JTC 21 on the Commission's request; each one grants presumption of conformity for the article it covers once the Commission cites it in the Official Journal. One is cited; the rest are drafts.

The harmonised standards under the EU AI Act, and where each stands
StandardCoversStatus
EN 18286:2026 — Quality management system for EU AI Act regulatory purposesArticle 17(1), and the technical documentation duty in Article 11(1), first sentencePublished July 2026; cited in the Official Journal 14 October 2026
prEN 18228 — AI risk managementArticle 9Draft; rejected at enquiry July 2026; expected late 2026
prEN 18229 series — Trustworthiness: logging (Part 1), human oversight, transparency, accuracy, robustnessArticles 12, 13, 14, 15Drafts; logging rejected at enquiry August 2026; oversight at enquiry October 2026; the rest 2027
prEN 18282 — Cybersecurity specifications for AI systemsArticle 15(5)Draft; rejected at enquiry July 2026; expected 2027
prEN 18283 — Managing bias · prEN 18284 — Quality and governance of datasetsArticle 10Drafts; expected 2027
prEN 18285 — Conformity assessment frameworkArticle 43, Annexes VI and VIIDraft; expected 2027

ISO/IEC 42001 was adopted in Europe as EN ISO/IEC 42001:2026 outside this request; adoption is not citation, and 42001 confers no presumption of conformity.

Risk management

NIST AI RMF

Govern, Map, Measure, Manage. Voluntary, but no longer without legal weight: Texas gives an affirmative defence to organizations substantially complying with NIST’s Generative AI Profile, or another recognised AI risk management framework. That second route reaches the ISO set too.

How we use it

ISO/IEC 23894:2023

AI-specific risk management guidance built on the ISO 31000 clause structure. Cannot be implemented on its own.

ISO 31000:2018

General risk management. Its stated purpose, creating and protecting value, is why risk and value belong in one conversation.

AI management systems

ISO/IEC 42001:2023

The certifiable AI management system standard, and the anchor for an auditable practice.

ISO/IEC 42005:2025

AI system impact assessment. Published, not proposed, and the method behind a risk register that addresses consequences.

ISO/IEC 42006:2025

Requirements for bodies auditing and certifying an AIMS. Determines what an auditor will actually look for.

AI security and privacy

ISO/IEC 27090:2026

Addressing security threats and compromises to AI systems. Guidance from the committee that writes ISO/IEC 27001: the first standardised catalogue of attacks on AI systems (data poisoning, evasion, membership inference, model extraction, model inversion, prompt injection and more), each with detection and mitigation, and how mitigations interact and decay across the life cycle. Builds on 27001 and 27002 and sits beside 42001. Not certifiable, and not a harmonised standard.

ISO/IEC 27090 in an ISMS, and in a 27001 internal audit

ISO/IEC 27091

Privacy protection for AI systems. Under development, at final draft stage.

Board governance

ISO/IEC 38507:2022

Written for the governing body: AI risk appetite, accountability, and a maintained view of every system in use.

Principles and ethics

OECD AI Principles

The shared intergovernmental baseline for trustworthy AI, adopted across dozens of countries and echoed in most national policy.

IEEE 7000-2021

Value-based engineering: ethical values elicited and traced into system design as requirements. Ethical value, never financial.

02

The three that carry most of the weight

Permalink to “The three that carry most of the weight”

For most enterprises, three instruments do the heavy lifting, and they interlock. The EU AI Act says what is required, NIST gives the structure to manage it, and ISO/IEC 42001 makes the result certifiable.

EU AI Act — the regulation with teeth

The first comprehensive horizontal AI law, structured by risk tier: prohibited practices, high-risk systems, limited-risk transparency duties, and minimal risk. Penalties reach up to 7% of global turnover for the most serious breaches. The timeline is staggered: prohibited practices and AI literacy are in force. General-purpose AI obligations applied from August 2025 and became enforceable in August 2026, alongside the Article 50 transparency duties. High-risk obligations were deferred to December 2027 for stand-alone systems and August 2028 where AI is embedded in regulated products. The practical task is classifying each system into the right tier, and establishing whether you act as provider, deployer, or both.

How it lands in your management system

NIST AI RMF — a working structure for AI risk

A voluntary US framework organized around four functions. Govern sets culture and accountability, Map establishes context and identifies risk, Measure analyses and tracks it, and Manage acts on it. It is not certifiable and creates no obligation of its own, which is precisely why it travels well: it gives a team somewhere to start without committing the organization to an audit. It also maps cleanly onto the controls in ISO/IEC 42001, so work done here is rarely wasted later.

How we use it

ISO/IEC 42001 — the AI management system

The first certifiable management-system standard for AI. It defines how an organization governs AI across its lifecycle: policy, risk assessment, controls, roles, and continual improvement. It has the same shape ISO/IEC 27001 gives information security, so an organization already certified there recognises the machinery. Companion standards complete it: 42005 for impact assessment and 42006 for the bodies that certify. For most enterprises this is the anchor, because it converts intent into a practice a third party can examine.

The standard, and where it sits under EN 18286

EN 18286 — the Article 17 quality management system

EN 18286:2026 is a quality management system standard for providers of high-risk AI systems, written for the EU AI Act and built around the AI system lifecycle. Quality here means conformity with the requirements that protect health, safety and fundamental rights. Its clauses cover the regulatory compliance strategy, design and validation, data, technical documentation, placing on the market, modifications, post-market monitoring and serious incidents.

Cited in the Official Journal on 14 October 2026 by Commission Implementing Decision (EU) 2026/2274, it gives presumption of conformity for the Article 17 requirements it covers: Article 17(1)(a) to (m) and the first sentence of Article 11(1). Nothing else. Applying the standard gives the presumption; a certificate does not, and third-party assessment is required only in the Annex VII cases.

Since October 2026 the harmonised standard for the Article 17 quality management system is EN 18286, which the Commission describes as a layer on top of ISO 9001 or ISO/IEC 42001 rather than a replacement. An existing ISO 9001 or ISO/IEC 42001 system gives the management chassis; on its own it establishes no conformity with Article 17 and carries no presumption of it. EN 18286, for the Article 17 requirements it covers, does provide that presumption. What the existing management system does not address is the regulatory compliance strategy, serious-incident reporting, communication with authorities and notified bodies, per-system technical documentation and post-market monitoring that Article 17 names.

It reaches providers of high-risk AI systems, from 2 December 2027 for Annex III systems and from 2 August 2028 for Annex I systems. A company using AI in its own processes is a deployer; Article 17 does not reach it.

The AI-specific instruments are recent. The disciplines they extend are not. Eight standards supply the foundation, and recognising them reduces the work. An organization with a mature risk function or an established governance practice already holds most of the machinery.

  • 01

    ISO 31000 — risk management

    Principles, framework, and process for managing risk in any organization. Its stated purpose is the creation and protection of value, which is why a serious AI risk practice also has to talk about return.

  • 02

    ISO/IEC 23894 — AI risk management

    The bridge between the two. It mirrors ISO 31000’s structure and extends its principles where AI differs: dynamism, the quality of available information, human and cultural factors. It presupposes the ISO 31000 baseline rather than replacing it.

  • 03

    ISO/IEC 38507 — governance for the governing body

    Board-level guidance covering AI risk appetite, accountability, and the maintained inventory of systems in use. It is the standard that makes AI oversight a governance responsibility rather than a technical one.

  • 04

    IEEE 7000 — ethics into design

    A method for eliciting ethical values from stakeholders and tracing them into system requirements. Developed largely in Vienna with supporting German standards work, it gives responsible AI an engineering process rather than a statement of principle.

  • 05

    ISO 9001:2026

    The quality management system standard. The sixth edition was published on 16 September 2026; 2015 certificates stay valid until 30 September 2029, and new certifications are issued only to the 2026 edition from 31 March 2028. It carries no AI-specific requirements. Applied here for the design control, verification, release and corrective-action discipline an AI project needs, for integrating an AI management system into an organization that already runs one and, for providers of high-risk AI systems, as the chassis EN 18286 is layered on.

  • 06

    ISO 21502:2020

    Guidance on project management. Named as one of the instruments this work is written against.

  • 07

    ISO 10005:2018

    Guidelines for quality plans. The shape a quality plan follows: what is being produced, how it will be verified, by whom, and what acceptance means.

  • 08

    ISO 22301 — business continuity

    The certifiable management system for business continuity, written to the same harmonised structure as ISO/IEC 42001. Business impact analysis, continuity strategies, plans and exercising live here and nowhere in the AI standard. So when an AI-dependent operation is disrupted, an organization already running ISO 22301 extends a discipline it has rather than improvising one.

ISO/IEC 27090 — security threats to AI systems

Published in October 2026 by the committee that writes ISO/IEC 27001, it is the first standardised catalogue of attacks on AI systems, each with how to detect and mitigate it. Three of its ideas are worth a buyer's time. Conventional controls can all be in place and the attack still succeed, because model extraction and poisoning ride on legitimate queries and data. Retraining turns operational data into an attack surface. And mitigations conflict and decay, so they are managed across the life cycle rather than installed once.

How we work with it: inside an ISMS, models, datasets and pipelines become assets, the catalogue feeds the risk register, and the mitigations become controls in the Statement of Applicability. In an ISO/IEC 27001 internal audit we assess the AI security controls against this guidance. It is guidance, not a certifiable standard, and not a harmonised standard under the AI Act.

04

Where each framework draws the line

Permalink to “Where each framework draws the line”

The instruments above operate at two levels, and the level decides how each one is used. Governance is the system by which the use of technology is directed and controlled: the governing body evaluates, directs and monitors. Management plans, builds, runs and monitors the activities delivering against that direction. Each body below describes both.

Where each framework draws the line — how each framework divides governance from management
FrameworkGovernance — the governing bodyManagement — the organization
ISOISO/IEC 38500, applied to AI by ISO/IEC 38507 — evaluate, direct, monitor. ISO/IEC 42001 reaches into this half too: policy, objectives, authorities and risk criteria (cl. 5, 6.1) and management review (9.3).ISO/IEC 42001 — plan, do, check, act. Support and operation (cl. 7, 8), and most of Annex A.
NIST AI RMFGOVERN — culture, accountability, risk tolerance. Cross-cutting by design rather than a stage.MAP · MEASURE · MANAGE — establish context, analyse, and act.
OCEGLEARN · ALIGN — context and stakeholders, then strategy to objectives.PERFORM · REVIEW — operate the controls, then test design and operating effectiveness.

Read across the table rather than down it. The three differ in vocabulary and in where they draw the line, not in whether the line exists. An organization already running one of these has the machinery for the others. What it needs is the AI-specific extension.

The ISO row is the one worth reading closely, because the split runs inside a single standard rather than between two. ISO/IEC 42001 carries directive content. At the two points where a decision belongs to the governing body and not to top management, its own notes hand off. Clause 5.2 refers AI policy development to ISO/IEC 38507, and clause 6.1.1 refers the determination of risk appetite to ISO/IEC 38507 and ISO/IEC 23894. The boundary sits in the notes, not in the requirements, which is where it is easiest to pass over.

ISO/IEC 42001 says as much itself. Its introduction states that the document contains no specific guidance on management processes. It adds that an organization may combine generally recognised frameworks, other international standards, and its own experience to implement the processes it needs. The standards themselves invite the comparison.

The Digital Omnibus on AI deferred parts of the Act. What moved is when the high-risk obligations begin to apply, not what they require. Risk management, data governance, logging, transparency and human oversight were not rewritten in substance, and they take longer to design and build than the new runway allows.

Deferred by the Digital Omnibus
  • —High-risk obligations for stand-alone Annex III systems (employment, credit, education, essential services) moved to 2 December 2027.
  • —High-risk obligations for AI embedded in products already under EU product-safety regulation (Annex I) moved to 2 August 2028.
  • —The deferral did not travel alone. The Omnibus also added an Article 5 prohibition covering AI whose reasonably foreseeable output includes non-consensual intimate imagery or child sexual abuse material absent adequate safeguards, with a transition to 2 December 2026. It narrowed the Annex I product scope. And it consolidated enforcement over AI systems built on general-purpose models by the same provider.
Binding regardless
  • —Prohibited practices and AI-literacy duties: in force since 2 February 2025.
  • —General-purpose AI obligations: applied since 2 August 2025, enforceable by the Commission since 2 August 2026.
  • —Article 50 transparency (disclosure that a user is dealing with AI, and marking of generative and synthetic media): since 2 August 2026. Generative systems already on the market before then have until 2 December 2026 for the machine-readable marking under Article 50(2).
  • —The substance of the deferred duties: risk management, data governance, logging and human oversight. They take longer to design and build than the runway now allows.
2 Feb 2025
in force

Prohibited practices · AI literacy

In force. Banned use cases apply outright, and providers and deployers must take measures to support the development of AI literacy among the people operating AI on their behalf.

2 Aug 2025
in force

General-purpose AI obligations

Applied to GPAI models placed on the market from this date: technical documentation, published training-data summaries, EU copyright compliance, and information sharing with regulators and downstream deployers.

2 Aug 2026
live now

Article 50 transparency · GPAI enforcement

Transparency duties apply: disclosure that a user is dealing with AI, and marking of generative and synthetic media. Generative systems already on the market before this date have until 2 December 2026 to meet the machine-readable marking requirement under Article 50(2). The Commission’s enforcement powers over general-purpose AI models start on the same date.

2 Aug 2027
upcoming

Legacy GPAI models

Models placed on the market before August 2025 lose their grace period and must be brought into compliance.

2 Dec 2027
upcoming

High-risk — stand-alone (Annex III)

Employment, credit, education, and essential-services systems. Deferred from 2026, and the extra runway is time to build the inventory and classification now.

2 Aug 2028
upcoming

High-risk — embedded (Annex I)

AI embedded in products already covered by EU product-safety regulation, following the conformity-assessment route those products already use.

Which of these obligations attach to you depends on the role the Act assigns, set out below.

06

Provider or deployer, which are you?

Permalink to “Provider or deployer, which are you?”

The EU AI Act assigns duties by role, not by whether you built the model. An organization using an AI system under its own authority is a deployer. Buying software with AI inside and using it makes you one, whether or not the purchase was ever reviewed as AI. The split is simple to state: the provider proves the system was built right, and the deployer proves it is being used right. The duties listed below attach to high-risk systems and begin to apply on 2 December 2027 for Annex III systems and 2 August 2028 for those embedded in regulated products. A deployer of a bought general-purpose assistant outside the high-risk tier owes the AI-literacy duty and nothing on these lists. It owes the Article 50 disclosure duties only where it uses the system for deepfakes or publishes AI-generated text on matters of public interest.

Provider — built right
  • —Conformity assessment and CE marking before the system reaches the market.
  • —Technical documentation, a risk management system, and data governance across the lifecycle.
  • —Instructions for use detailed enough to make compliant deployment possible.
  • —Post-market monitoring and serious-incident reporting.
Deployer — used right
  • —Use in accordance with the provider’s instructions (Art. 26).
  • —Human oversight assigned to people with both the competence and the authority to intervene.
  • —Input data relevant and sufficiently representative, wherever you control it.
  • —Operation monitored, and use suspended where risk emerges.
  • —Automatically generated logs retained for at least six months.
  • —Affected persons informed, including workers and their representatives before a high-risk system goes live.
  • —A Fundamental Rights Impact Assessment where Article 27 applies: a deployer document no vendor can produce for you.
And the line moves

Article 25 turns a deployer into a provider. If you put your own name or trademark on a high-risk system, make a substantial modification to it, or repurpose it into a high-risk use it was not sold for, the full provider obligation set attaches. That includes conformity assessment, CE marking and technical documentation. Fine-tuning a bought model can cross that line. So can pointing a general-purpose tool at a hiring, credit, or access decision.

None of this is transferable by contract. A vendor’s assurances are evidence that you exercised diligence, but they are not a substitute for obligations the regulation places on you directly. Nor is the pattern EU-specific. ISO/IEC 42001 is scoped to organizations that provide or use AI, and the standard says both. It has a dedicated control set for responsible use (A.9) and for allocating responsibilities across third parties and suppliers (A.10).

The rules are supranational; enforcement is not. Which authority comes asking depends on where the organization sits, what the system decides, and which sector it operates in. In both regions we cover, that answer changed during 2026.

DACH

Germany’s implementing law is the Gesetz zur Marktüberwachung und Innovationsförderung von künstlicher Intelligenz (KI-MIG). It passed the Bundestag on 11 June 2026, cleared the Bundesrat on 10 July, and entered into force at the end of that month, ahead of the Act’s own 2 August 2026 date. It settles a question that had been open for a year: who comes asking.

Germany chose a hybrid model rather than a new agency. The Bundesnetzagentur becomes the central market-surveillance authority, the single point of contact toward the EU AI Office, and the central complaints office. A coordination and competence centre (§ 5 KI-MIG) and the KI-Marktüberwachungskammer (§ 4 KI-MIG) support it. Sector regulators supplement it where they already hold the mandate, notably BaFin for the financial sector. Cybersecurity of high-risk AI systems runs through the Cyber Resilience Act’s market-surveillance authority, with the BSI holding the role in the interim (§ 10 KI-MIG).

That split is not academic. Inside a single bank, systems can fall under different supervisory routes depending on what they decide, so establishing which authority governs which system is genuine work. For smaller organizations there is an upside worth using: the regulatory sandbox route is designed with access for SMEs and startups in mind.

Austria and Switzerland complete the picture differently. Where Germany distributes supervision across a hybrid, Austria concentrates it in one unit inside the telecoms regulator. Under its implementing law, the KI-Servicestelle at RTR has been the market-surveillance authority since August 2026, and it also serves as the national contact point and sandbox operator. Swiss firms sit outside the Act domestically but are caught by its extraterritorial reach the moment they place an AI system on the EU market or its output is used there.

Bundesnetzagentur (BNetzA)Central market surveillance · notifying authority · EU AI Office contact point · complaints
BaFinFinancial-sector AI supervision, within the existing prudential mandate
BSICybersecurity of high-risk AI systems, in the interim until the Cyber Resilience Act’s market-surveillance authority is designated (§ 10(4) KI-MIG)
BfDI and the LandesdatenschutzbehördenData-protection oversight: federal bodies to the BfDI, private-sector controllers to the Länder authorities, which are market-surveillance authorities for the law-enforcement and justice uses
Länder authoritiesAI operated by public bodies — education, police, justice, social services
DAkkSAccreditation of conformity-assessment bodies
Canada

Canada has no comprehensive AI statute. The Artificial Intelligence and Data Act died with Bill C-27 when Parliament was prorogued in January 2025 and has not been reintroduced. In June 2026 the federal government launched a national AI strategy instead, with privacy and online-safety modernization among its stated priorities.

The absence of a statute is not an absence of obligation. What binds a Canadian organization today is privacy law and sector supervision. The Privacy Commissioner reads PIPEDA broadly enough to reach most AI systems handling personal information. Quebec’s Law 25 requires disclosure where a decision is made exclusively by automated processing, and gives the individual a right to an explanation. That is the closest thing to AI-specific law in force in the country. Ahead of both, federally regulated financial institutions face OSFI’s revised Guideline E-23 on model risk management, which brings AI and machine-learning models expressly into scope for the first time. It takes effect 1 May 2027, which makes the preparation a 2026 exercise.

For organizations operating on both sides of the Atlantic, the binding constraint is often still the EU AI Act. Its extraterritorial reach attaches the moment a system is placed on the EU market or its output is used there, regardless of where the model was built.

Privacy Commissioner (OPC)PIPEDA — personal information handled by AI systems
OSFIGuideline E-23 — model risk management at federally regulated financial institutions
Commission d’accès à l’informationQuebec Law 25 — automated decision disclosure and explanation rights
Treasury Board SecretariatDirective on Automated Decision-Making — algorithmic impact assessment for federal institutions

A compliance programme built against one snapshot of this landscape can become outdated quickly. That is no reason to wait, but it makes more sense to build against obligations than against dates.

Four examples since early 2025. The Digital Omnibus deferred the EU AI Act’s high-risk deadlines after they had already been set. Colorado repealed and replaced its AI Act before the original ever took effect. Canada’s Artificial Intelligence and Data Act died with Bill C-27 and has not returned. And in Europe, the harmonised standards began to arrive: EN 18286 for the Article 17 quality management system was cited in the Official Journal on 14 October 2026, the first under the Act, with presumption of conformity for the Article 17 requirements it covers. The rest, including prEN 18228 for risk management, are still drafts; three were rejected at public enquiry over the summer and are expected in 2027. Until each is cited, conformity with its article cannot be claimed through it.

Nor is the landscape uniform. No jurisdiction outside the EU has a comparable timetable. The United States has no federal statute at all, only a state-by-state patchwork, with Texas and California in force since January 2026.

The practical consequence is a design principle. Build the risk management, data governance, logging and human oversight the regulation describes, and keep the mapping to specific clauses, standards and dates in a layer you can change. The obligations have proved far more stable than the deadlines and instruments attached to them. That is also why the classification work matters more than the calendar. Once each system’s tier, role and exposure are established, mapping them onto whichever regime applies is comparatively mechanical, and it survives the next amendment.

The map is not the answer. It is the start of one.

Read together, the ISO set forms a coherent spine: 38507 places accountability with the board, 23894 adapts risk management to AI, 42005 assesses impact, 42001 holds the management system, and 42006 defines what certification demands of it. Which of these apply to you, and what to do about them, depends on your systems, your risk surface and the role the regulation assigns you. A good place to start is a conversation about exactly that.

More questions

Why do the deadlines move when the obligations stay?

Because the dates and instruments around the obligations have proved far less stable than the obligations themselves. The Digital Omnibus on AI moved the application dates; the quality management duty in Article 17 itself was not amended.

Where does Canada stand on AI regulation?

Canada has no AI statute; privacy law and sector supervision apply. The Artificial Intelligence and Data Act died with Bill C-27 in January 2025 and was not reintroduced. The national strategy 'AI for All' was launched on 4 June 2026. OSFI Guideline E-23 on model risk management, with AI and machine learning expressly in scope, takes effect on 1 May 2027.

What does Switzerland require?

Switzerland has no AI-specific law; Swiss companies are caught only through the EU AI Act when they place systems on the EU market. The Federal Council has decided to ratify the Council of Europe AI Convention and is preparing a consultation draft due at the end of 2026.

Get a clear map of what applies to you.

Tell us where AI sits in your organization. We will respond with a read on which instruments apply, which are worth adopting voluntarily, and where one piece of work can satisfy several at once.