EU AI Act for automotive
Most vehicle AI arrives through the Annex I product route at 2 August 2028, assessed inside type approval rather than through a parallel procedure. The sharper question is in the cabin, where emotion inference meets an Article 5 prohibition framed around the workplace.
One vehicle, three possible dates
| System | Route | Date |
|---|---|---|
| ADAS, autonomous driving functions, safety-critical control | Annex I, safety component of a type-approved product | 2 Aug 2028 |
| In-cabin biometric identification: identifying the occupant rather than confirming a claimed identity | Annex III point 1(a) | 2 Dec 2027 |
| In-cabin emotion inference | Annex III point 1(c), or Article 5 prohibition if the cabin is a workplace | Art. 5 live since 2 Feb 2025 |
| Voice assistants, generative in-car features | Article 50 transparency: any tier | 2 Aug 2026 |
| Fleet driver scoring and task allocation | Annex III point 4(b): worker management | 2 Dec 2027 |
A connected vehicle with a voice assistant owes Article 50 today, whatever happens in 2027 and 2028.
Where this usually goes next
Three situations account for most people reading this page. Each has a different answer.
A deal is blocked on an AI questionnaire
Legal will not sign until you can evidence how AI is governed. HumanAudit’s AI Trust Package is a fixed $3,500 over five business days: a public trust page, a pre-filled SIG Lite / CAIQ / SSPA Section K questionnaire bank, and your AI inventory and classification.
You need ISO/IEC 42001 documentation
23 clause-mapped AIMS documents with all 38 Annex A controls pre-populated, editable and yours to keep, from $199. Or score your gaps first: 18 questions, free, no signup to begin.
You are not sure what reaches you
Twenty minutes with the founder. No prep, no deck, straight to the person accountable for the work. If none of this applies to you, you get told that on the call.
This reference is published by HumanAudit Inc. Not a law firm, not an accredited certification body, not a registered auditor. We build documentation, your counsel interprets it, and an accredited body of your choosing certifies you. How this is funded →
The workplace line, and why it is sharp
Article 5 prohibits inferring emotions of a natural person in the areas of workplace and education institutions, except where the system is intended to be put in place or into the market for medical or safety reasons.
Two questions follow that a vehicle manufacturer has to answer deliberately.
1. Is the cab a workplace?
For a professional driver, haulage, delivery, taxi, bus, the vehicle is where they work. On the ordinary meaning of the words that is a workplace, which brings in-cabin emotion inference into the Article 5 prohibition rather than the high-risk regime.
Expert analysis. The Regulation does not define “workplace” for this purpose and no guidance or enforcement decisions address it. A manufacturer selling the same platform into private and commercial vehicles should not assume one answer covers both.
2. Is it a safety system or a comfort feature?
Drowsiness and attention monitoring for road safety is a genuinely arguable safety case within the exception. Inferring mood, stress, frustration or engagement to adapt music, lighting or advertising is not a safety purpose, and the exception is drafted around purpose.
The practical consequence: the same sensor stack can sit either side of a prohibition depending on what you do with the inference. That is a product decision, and it should be documented as one before launch rather than reconstructed afterwards.
What runs alongside
- Type approval. Where AI is a safety component of a type-approved vehicle, the AI Act requirements are assessed within the existing conformity procedure. Your technical service and approval authority relationships carry over; the AI-specific evidence does not exist yet.
- ISO 26262 and SOTIF transfer partly. Functional safety and safety-of-the-intended-functionality practice give substantial Article 9 and Article 15 evidence, oriented to hazard, not to fundamental rights, which Article 9 also requires. Article 9 →
- Article 10 bias examination has no automotive analogue. Examining whether perception performance differs across groups of persons on whom the system is used is an AI Act requirement, not a functional safety one. Article 10 →
- Fleet operators are deployers. A haulier deploying driver scoring owes Article 26 duties and Article 26(7) worker notification, separately from anything the manufacturer owes. Roles →
Status labels on this page
Verified fact: The Art. 6(1) Annex I route and its 2 August 2028 date; Annex III point 1 sub-points; the Art. 5 prohibition on emotion inference in the workplace and its medical or safety exception; Art. 50 application from 2 August 2026; and Annex III point 4(b).
Expert analysis: The workplace analysis for commercial vehicle cabs, the safety-versus-comfort distinction, and the assessment of what ISO 26262 and SOTIF practice transfers.
Unsettled: Whether a commercial vehicle cab constitutes a workplace for Article 5 purposes. No definition, guidance or enforcement decisions address it.
Decide the in-cabin position before launch
The emotion-inference question is a product decision with a prohibition on one side of it. Document which purpose the inference serves, who decided, and on what basis, before the platform ships into commercial fleets.
Not sure where you sit?
The classifier maps your system against Articles 5, 6, 50 and Annex III. Twelve questions, no email.
Frequently asked
Does the EU AI Act apply to cars?
AI that is a safety component of, or is itself, a product covered by the Annex I harmonisation legislation and required to undergo third-party conformity assessment follows the Article 6(1) route and applies from 2 August 2028. Vehicle type-approval legislation sits within Annex I. Separately, in-cabin systems that identify people biometrically, categorise them or infer emotion can engage Annex III point 1 or the Article 5 prohibitions, on different dates.
Is driver monitoring prohibited under the EU AI Act?
It depends on what is inferred and where. Article 5 prohibits inferring emotions of a natural person in the areas of workplace and education institutions, except where the system is intended to be put in place or into the market for medical or safety reasons. Drowsiness and attention monitoring for road safety is a genuinely arguable safety case. Inferring a driver's mood, stress or engagement for non-safety purposes in a commercial vehicle, which is that driver's workplace, is a materially weaker position.
Which date applies to automotive AI?
Usually 2 August 2028, through the Article 6(1) Annex I product route, following Regulation (EU) 2026/1744. But a vehicle can carry systems on more than one date: an in-cabin biometric identification system would be assessed under Annex III point 1 at 2 December 2027, and any AI feature that interacts with occupants or generates synthetic content owes Article 50 transparency, which has applied since 2 August 2026.