Skip to content
SimpleAct Logo

Guide · Executives & C-level

As of 7 October 2026 · after the Digital Omnibus

EU AI Act for Executives and C-Level: Duties, Deadlines and Liability Questions

The AI Act addresses the organisation – and therefore the leadership that runs it. This guide shows, without alarmism, what executives and boards need to know and decide: which duties already apply, what fines realistically mean, how to think about liability questions, and which tasks are due in the next 90 days.

  • Duties with article and deadline
  • Fines put in perspective
  • RACI excerpt for leadership
  • Checklist with progress
app.simpleact.de

Dashboard

Demo Organisation GmbH · overall status

A

4

AI systems

1

high-risk

1

review overdue

87%

audit readiness

Open items and deadlines

  • Application screening: human oversightArt. 14 · owner Dr. Anna SchmidtOpen
  • Service chatbot “Lia”: approvalGovernance · reviewer L. WeberIn review
  • Website translation pluginAI inventory · not yet classifiedUnassessed
  • High-risk duties Annex IIIDeadline 2 Dec 2027 · 1 system affectedDeadline

Report for management, auditors and customers

  • PDF
  • DOCX
  • incl. audit history

Sample data · audit readiness score from Professional

In short

Executives do not have to implement the AI Act themselves, but they have to organise it: appoint owners, set up an AI inventory and an approval process, commission AI literacy measures and ask for regular reports. Fines are imposed on the organisation; how far personal responsibility reaches depends on national law and the individual case.

Basics

Why the AI Act is a leadership topic

The AI Act (Regulation (EU) 2024/1689) assigns duties to roles: providers, who develop AI systems and place them on the market under their own name, and deployers, who use AI systems under their own authority (Art. 3(3) and (4)). Almost every organisation is a deployer today – through chat assistants, AI features in office software, translation services or recruiting tools. The duties fall on the organisation as a whole. Organising how they are met is therefore a job for leadership.

Some duties already apply. Since 2 February 2025 certain AI practices are prohibited (Art. 5), and providers and deployers must take measures on the AI literacy of their staff (Art. 4). Since 2 August 2026 the transparency duties in Art. 50 apply, for example to chatbots and AI-generated content. The Digital Omnibus (Regulation (EU) 2026/1744, in force since 27 July 2026) rewrote Art. 4: it now requires measures to promote AI literacy rather than ensuring a particular level.

The most extensive duties concern high-risk AI – systems in the areas listed in Annex III, such as candidate selection, employee performance assessment, creditworthiness checks or access to essential services. After the Digital Omnibus these duties apply from 2 December 2027, and for AI in regulated products under Annex I from 2 August 2028. That sounds far off, but it is tight for building an inventory, roles and evidence if you only start in 2027.

For leadership it comes down to three questions. First: do we know which AI we use and for what? Second: is it clear who is responsible for what and who decides? Third: can we prove it – to supervisory authorities, customers, auditors and our own supervisory board? An organisation that can answer yes has done most of the work before the strictest duties kick in.

This guide deliberately avoids horror scenarios. The maximum fines are real, but they are ceilings. For most organisations the bigger risk is a different one: projects stopped because questions about classification, the works council or data protection were asked too late, and customer questionnaires that nobody can answer with confidence.

prohibitions (Art. 5) and AI literacy (Art. 4) apply
2 Feb 2025
prohibitions (Art. 5) and AI literacy (Art. 4) apply
high-risk duties for Annex III systems after the Digital Omnibus
2 Dec 2027
high-risk duties for Annex III systems after the Digital Omnibus
questions leadership should be able to answer
3
questions leadership should be able to answer

Duties

What the EU AI Act means for executives

The table shows the duties relevant to a typical organisation acting as deployer, and what leadership should set in motion for each. The duty falls on the organisation; the right-hand column describes the leadership task behind it.

Scroll the table sideways →

What the EU AI Act means for executives
DutyBasisAddresseeApplies since / fromLeadership task
Promote AI literacyArt. 4 AI Act (amended)Providers and deployers2 Feb 2025Commission training and briefings per user group, approve budget, ask for evidence.
Rule out prohibited practicesArt. 5 AI ActAll operators2 Feb 2025A clear rule in the AI policy that prohibited practices are not used; check in the approval process.
New prohibitions (Omnibus)Art. 5 AI Act (amended)All operators2 Dec 2026Non-consensual intimate deepfakes of identifiable people and child sexual abuse material – mainly relevant to providers of generative systems.
Transparency for chatbots and AI contentArt. 50 AI ActProviders and deployers2 Aug 2026Make sure customers can tell when they interact with AI and when content is AI-generated.
High-risk AI: use, oversight, monitoringArt. 26(1), (2), (5) AI ActDeployers of high-risk AI2 Dec 2027Name an owner and competent people for human oversight for each high-risk system; provide resources.
Inform workers beforehandArt. 26(7) AI ActEmployers as deployers2 Dec 2027Inform workers’ representatives and affected workers before a high-risk system is used at the workplace.
Inform affected personsArt. 26(11) AI ActDeployers (Annex III)2 Dec 2027Set up processes so people about whom a high-risk system helps decide are informed.
Fundamental rights impact assessmentArt. 27 AI ActPublic bodies, private providers of public services, Annex III 5(b)/(c)2 Dec 2027Have it checked whether the organisation is covered – for example for creditworthiness or life and health insurance.
Avoid or steer a switch to the provider roleArt. 25 AI ActDeployers, distributors, third parties2 Dec 2027 (Annex III)Decide whether to offer own products with AI – then provider duties apply, with much more effort.
Know the fine frameworkArt. 99 AI ActAll operatorssince 2 Aug 2025Put the risk in perspective and include it in enterprise risk management.

Simplified overview for deployers. Organisations that provide AI systems themselves have additional duties such as risk and quality management (Art. 9, 17). High-risk duties apply only to Annex III or Annex I systems.

In detail

Six topics that belong on leadership’s desk

Each topic covers what the Regulation requires, what that means for leadership decisions and which recommendations have proven themselves in practice.

01

Organise responsibility, don’t implement it yourself

Art. 4, 26(2) AI Act

The AI Act does not prescribe a particular organisation. It requires results: literacy measures, human oversight by people with the necessary competence, training and authority (Art. 26(2)), monitoring, evidence. How the organisation gets there is leadership’s call. A simple model has proven itself: a named person for AI governance with a clear mandate and time budget, system owners in the business units and a short approval process for new use cases.

Leadership keeps three tasks that cannot be delegated: grant the mandate and provide resources, adopt the AI policy, and decide borderline cases and high-risk approvals. Everything else can – and should – sit with the specialists. Executives who approve every tool become a bottleneck; executives who decide nothing end up without a line.

  • Named person with mandate and time budget
  • AI policy adopted by leadership
  • High-risk and borderline cases stay with leadership
02

Liability questions for management – a sober view

Art. 99 AI Act · national law

Fines under the AI Act are imposed on the operators that breach a duty – usually the organisation. The Regulation contains no provision of its own on the personal liability of managing directors or board members. Whether and to what extent members of management can be held personally liable depends on the applicable national law and the circumstances of the case. Clarify this with your legal advisers, not on the basis of marketing statements – ours included.

Whatever the legal assessment, a general principle of good corporate governance applies: leadership provides an organisation in which duties are identified, owners are named and deviations are noticed. For AI that means an inventory, a policy, an approval process, training and regular reports. This organisation can be documented – and that documentation is exactly what supervisory boards, auditors and authorities ask for.

Recommendation: ask for a report on the state of AI governance at least once a year and keep it on record. Check with your insurer how AI-related risks are treated in existing policies. And avoid the opposite of governance: AI projects that go live without classification and without data protection and the works council being involved.

  • Fines are usually imposed on the organisation
  • Personal responsibility: national law, individual case, legal advice
  • Document the organisation, ask for annual reports
03

Putting fines in perspective

Art. 99 AI Act

The Regulation tiers the maximum amounts. Breaches of the Art. 5 prohibitions can be fined up to €35 million or 7% of worldwide annual turnover. Breaches of a range of other duties – including the deployer duties in Art. 26 and the transparency duties in Art. 50 – up to €15 million or 3%. Supplying incorrect, incomplete or misleading information to authorities up to €7.5 million or 1%. The higher amount applies; for SMEs and start-ups the lower one (Art. 99(6)). Since the Digital Omnibus, this also applies to small mid-caps (Art. 3(14b)), but only for fines under paragraphs 4 and 5 (Art. 99(6a)).

Two points are often overlooked. First, these are maximum amounts, not standard amounts. Second, the AI literacy duty in Art. 4 has no fine tier of its own in Art. 99. That does not make it unimportant: an organisation without sufficiently competent staff will struggle to evidence human oversight under Art. 26.

  • Art. 5: up to €35M / 7%
  • Art. 16, 22–26, 31, 33, 34, 50: up to €15M / 3%
  • Incorrect information: up to €7.5M / 1%; SMEs the lower amount
04

Deadlines and what they mean for planning

Art. 113 AI Act · Regulation (EU) 2026/1744

The Digital Omnibus moved the high-risk deadlines: Annex III from 2 December 2027 (previously 2 August 2026), Annex I from 2 August 2028 (previously 2 August 2027). For own generative systems placed on the market before 2 August 2026, the marking duty under Art. 50(2) applies from 2 December 2026. Two new prohibitions take effect on the same date.

For planning that means: the time until the end of 2027 is enough for a solid approach if you start now – not if you start in autumn 2027. In practice the documents do not take longest; the alignment does: with business units on purposes, with vendors on documentation, with the works council on use at the workplace. The Omnibus also brings relief for SMEs and small mid-caps, mainly for provider duties such as technical documentation and quality management.

  • Now: meet Art. 4, 5 and 50
  • 2026/2027: inventory, classification, roles, vendor documentation
  • By 2 Dec 2027: high-risk duties evidenced
05

Customers, purchasing and tenders

Market practice (recommendation)

Long before an authority asks, customers do. Supplier questionnaires and tenders increasingly ask: which AI systems do you use? How are they classified? Who is responsible? Is there an AI policy? A team that can answer with an orderly report shortens the sales cycle; a team that first has to send an email round through the departments loses time and credibility.

Watch the wording: credible answers describe the implementation status – inventory, classification with reasons, owners, measures. Blanket promises such as “fully compliant” are hard to keep under a regulation whose high-risk duties only apply from December 2027 and whose guidelines are partly still in draft.

  • Prepare a standard answer for supplier questionnaires
  • Describe the status rather than promise compliance
  • Keep the report exportable
06

Works council and employees

§ 80(3), § 87(1) no. 6, § 90(1) no. 3 BetrVG · Art. 26(7) AI Act

In German organisations with a works council, early involvement often sets the pace. An employer planning to use AI in work processes and workflows must inform the works council in good time (§ 90(1) no. 3 BetrVG). Where a system is suitable for monitoring employees’ behaviour or performance, a co-determination right can exist (§ 87(1) no. 6 BetrVG). Where the works council has to assess the introduction or use of AI, calling in an expert is deemed necessary to that extent (§ 80(3) BetrVG).

For high-risk AI at work the AI Act adds a duty: workers’ representatives and affected workers must be informed before the system is put into service (Art. 26(7)). Recommendation: a framework agreement on AI that sets ground rules and adds individual tools through a simplified procedure. That saves negotiating each new tool separately.

  • Inform early (§ 90 BetrVG)
  • Check co-determination where monitoring is possible (§ 87 BetrVG)
  • Framework agreement instead of case-by-case (recommendation)

RACI excerpt

Where management decides – and where it doesn’t

Excerpt from an AI governance RACI matrix with the tasks leadership takes part in. The highlighted column shows the role of management. Each task has exactly one A. The assignment is a suggestion for a mid-sized organisation.

Scroll the table sideways →

Where management decides – and where it doesn’t
TaskManagementAI governance leadBusiness unit / system ownerIT lead / CTOData protection (DPO)Legal
Set mandate, resources and AI governance leadOrganisational decisionA/RCIIII
Adopt the AI policyArt. 4, 5 AI ActARCCCC
AI literacy: approve the training planArt. 4 AI ActARCCCI
Approve high-risk and borderline casesArt. 6, 26 AI ActARCCCC
Approve non-critical standard toolsInternal processIARCC–
Inform workers and works councilArt. 26(7) AI Act · § 90 BetrVGARC–IC
Decide on serious incidentsArt. 26(5) AI ActARRRCC
Annual governance reportRecommendationARCCCC
  • RResponsible – does the work
  • AAccountable – owns and decides
  • CConsulted – asked beforehand
  • IInformed – kept informed
  • Leadership does not need to approve non-critical standard tools itself. Delegating to the AI governance lead with clear criteria keeps the process fast.
  • DPOs inform, advise and monitor (Art. 39 GDPR). They should not grant approvals, because they would then be monitoring their own decisions (Art. 38(6) GDPR).
  • The full matrix with twelve tasks including the works council is in the AI governance framework.

Interactive

Leadership’s top tasks for the next 90 days

Twelve tasks in three phases – most of them decisions, not work packages. Tick off what is done; expand any task for details.

0 of 12 tasks done

0%

Your progress is stored in this browser only, never transmitted.

Days 1–30

0/4

Mandate and overview

  • Name a person responsible for AI governance
  • Commission a first AI inventory
  • Announce an interim rule: no new AI tools without a request
  • Inform the works council about AI plans early

Days 31–60

0/4

Set the rules

  • Adopt the AI policy
  • Define the approval process and delegation limits
  • Approve a training plan per user group
  • Have potential high-risk cases prioritised

Days 61–90

0/4

Report and follow up

  • Approve a standard answer for customer and supplier questionnaires
  • Add AI risks to enterprise risk management
  • Receive the first governance report
  • Set a reporting rhythm

Order and time frame are a recommendation, not a legal requirement. With many high-risk cases or own AI products it takes longer.

From leadership’s point of view

Questions executives and boards ask themselves

And answers without legalese. They do not replace legal advice, but they help you ask specialists the right questions.

“Does the AI Act even apply to us? We don’t develop AI.”

Very likely yes – as a deployer. Anyone using chat assistants, AI features in office software or AI in recruiting already has duties under Art. 4, 5 and 50. The extensive high-risk duties only concern certain purposes.

Understand risk classes

“What will this cost us?”

For a mid-sized organisation without its own AI products, mostly internal time: inventory, policy, training, approval process. Effort rises sharply with high-risk systems or own AI products. A tool can replace spreadsheets and email rounds; SimpleAct’s prices are listed openly.

See pricing

“Am I personally liable?”

Fines under the AI Act are usually imposed on the organisation. Whether and how far members of management are personally responsible depends on national law and the individual case – that belongs in a conversation with your legal advisers. A documented, working organisation is the best foundation either way.

“Can we wait with AI projects until 2027?”

For the high-risk duties yes, for the rest no. Prohibitions, AI literacy and transparency duties already apply. And anyone who wants to run high-risk systems with evidence in 2027 needs inventory, classification, vendor documentation and roles beforehand.

All deadlines

“Who should do this here?”

A named person with mandate and time, supported by IT, data protection and legal. In smaller organisations often the IT lead or the person who also coordinates data protection. The mandate from leadership matters, not the title.

AI governance framework

“What do we tell our customers?”

Describe the implementation status: which AI systems are in use, how they were classified, who is responsible and which measures exist. An exportable report is more convincing than a promise of compliance.

Deadlines

The dates leadership should know

These dates come from the AI Act and the Digital Omnibus. Next to each is the decision it calls for.

  1. 2 Feb 2025

    Prohibited practices (Art. 5) and AI literacy (Art. 4)

    AI policy with prohibitions, approve the training plan.

  2. 2 Aug 2025

    Duties for general-purpose AI models, governance, penalties

    The fine framework applies; relevant for risk management.

  3. 2 Aug 2026

    General application, transparency duties under Art. 50

    Customer contact with AI must be recognisable.

  4. 2 Dec 2026

    New prohibitions (Omnibus), marking for legacy systems

    Mainly relevant for own generative offerings.

  5. 2 Dec 2027

    High-risk duties for Annex III systems

    Owners, oversight, informing workers and affected persons must be in place.

  6. 2 Aug 2028

    High-risk duties for Annex I systems

    Relevant for manufacturers of regulated products with AI components.

Status after the Digital Omnibus, Regulation (EU) 2026/1744, in force since 27 July 2026. All AI Act deadlines →

With SimpleAct

The status at a glance – not from five spreadsheets

SimpleAct shows leadership how many AI systems there are, how they are classified, what is open and where a deadline has passed. The view on the right shows the Governance Hub with sample data from the demo tenant.

  • Dashboard

    all plans

    Total systems, risk classes, open assessments, gaps and recent activity.

  • Governance workflows

    Professional

    Owner, reviewer, due date and approval per object; overdue items become visible. Plus an audit readiness score.

  • Reports

    all plans

    Reporting as PDF or DOCX with versioning – for management, the supervisory board, auditors or customers. Register export from Professional.

  • Fast start

    Pilot project

    In a pilot project, six AI systems – one of them high-risk – were captured and assessed in around four hours.

Prices per the pricing page, plus VAT: Starter €199 per month (€159 billed annually), Professional €349 (€279 annually), Enterprise on request. Free 7-day trial (credit card required, cancel during the trial).

app.simpleact.de

Governance Hub

Owners, reviews and evidence per object

A
Workflows

12

In review

3

Overdue

1

Evidence

28

RISK_ASSESSMENT

AI-based Application Screening

FINAL
OwnerReviewerDue dateApproval

ANNEX_IV_DOSSIER

Service Chatbot "Lia"

IN REVIEW
OwnerReviewerDue dateApproval

COMPLIANCE_CHECKLIST

Marketing Copy Generator

DRAFT
OwnerReviewerDue dateApproval

FAQ

Frequently asked questions from executives and the C-suite

The AI Act’s duties fall on the organisation; leadership organises how they are met: appoint owners, provide resources, adopt an AI policy, commission AI literacy measures, decide high-risk and borderline cases and ask for regular reports.
Since 2 February 2025 the Art. 5 prohibitions and the duty to take AI literacy measures under Art. 4; since 2 August 2026 the transparency duties under Art. 50. After the Digital Omnibus the high-risk duties apply from 2 December 2027 for Annex III systems and from 2 August 2028 for Annex I.
Under Art. 99 up to €35 million or 7% of worldwide annual turnover for prohibited practices, up to €15 million or 3% for breaches of other duties such as Art. 26 and Art. 50, up to €7.5 million or 1% for incorrect information to authorities – the higher amount in each case, for SMEs the lower (Art. 99(6)), for small mid-caps only under paragraphs 4 and 5 (Art. 99(6a)). Art. 4 has no fine tier of its own.
The AI Act contains no rule of its own on the personal liability of members of management; its fines are usually imposed on the organisation. Whether personal responsibility exists depends on national law and the individual case and should be clarified with legal advisers. A documented governance organisation is the best foundation either way.
Art. 4 requires measures for staff and other persons dealing with AI systems on your behalf. Leadership should be informed enough to assess the policy, approvals and reports. A short briefing on risk classes, duties and deadlines is a proven way to get there.
The AI Act does not require an AI officer. In practice a named person for AI governance with mandate, time budget and a reporting line to leadership has proven itself. Clear responsibility matters, not the title.
It moves the high-risk deadlines (Annex III to 2 December 2027, Annex I to 2 August 2028), rewrites Art. 4 as a duty to take promotion measures, adds two prohibitions from 2 December 2026 and brings relief for SMEs and small mid-caps, mainly for provider duties. Regulation (EU) 2026/1744 has been in force since 27 July 2026.
The deployer duties under Art. 4, 5 and 50 apply regardless of size. For fines, SMEs face the lower of the two amounts (Art. 99(6)); since the Omnibus, small mid-caps do too, but only for fines under paragraphs 4 and 5 (Art. 99(6a)). The Digital Omnibus eases mainly provider duties for SMEs and small mid-caps, for example through simplified technical documentation.
In Germany, an employer planning to use AI in workflows must inform the works council in good time (§ 90(1) no. 3 BetrVG). For systems suitable for monitoring behaviour or performance a co-determination right can exist (§ 87(1) no. 6 BetrVG). For high-risk AI at work, Art. 26(7) AI Act requires informing workers’ representatives beforehand.
A first inventory with provisional classification is usually possible within a few weeks. With SimpleAct, a pilot project captured and assessed six AI systems – one of them high-risk – in around four hours. You can also try it for 7 days (credit card required).

More questions? We're happy to help. Send email · Try 7 days (credit card required)

Sources and status

Last reviewed on · SimpleAct editorial team

This page is for orientation and is not legal advice, in particular not on liability questions. The text published in the Official Journal of the EU is authoritative. RACI assignment, checklist and time frames are recommendations, not legal requirements.

Changes to this page

  • 7 October 2026 – Fines: lower amount for small mid-caps (Art. 99(6a)) added.
  • 6 October 2026 – Rebuilt as a guide: duties table with articles, liability questions in general terms, fines under Art. 99, RACI excerpt, interactive 90-day checklist, deadlines after the Digital Omnibus, ten FAQs. Blanket statement on personal liability removed.
  • 18 May 2026 – First version as a role page for executives.

The state of your AI governance in one report

SimpleAct connects the AI inventory, classification, approvals, deadlines and evidence – so leadership decides on facts.