Nemko Digital Insights

15 EU Regulations Shaping AI-embedded Products | Nemko Digital

Written by Mónica Fernández Peñalver | September 11, 2025

"With great power comes great responsibility." For companies embedding AI into physical products, that responsibility takes the form of legal compliance. In the EU, it is not one law but a dense web of regulations that together demand proof of safety, security, and accountability.

We see many companies are still far from ready: some focus too narrowly on the AI Act, others underestimate the organizational and technical changes required. The result can be delayed market access, loss of competitiveness, or liability exposure.

 

15 EU Regulations Shaping AI-embedded Products

AI-embedded products are not governed by a single law, but by a network of 15 EU regulations that span data, cybersecurity, product safety, market rules, and health-specific frameworks. Together, they define the compliance landscape:

AI and Data

  • General Data Protection Regulation (GDPR): Governs personal data processing, essential when AI relies on sensitive user or health data.
    Applicability: In force (full text)
  • Data Act: Ensures fair access and use of data from connected devices for users, third parties and governments.
    Applicability: In force (full text)
  • Data Governance Act: Establishes trusted frameworks for data sharing and intermediaries, particularly in contexts like public-sector data reuse, data altruism (sharing for public good), and neutral data intermediation services.
    Applicability: In force (full text)
  • ePrivacy Regulation (in draft): expected to update privacy rules for digital communications.
    Applicability: In draft / TBC
  • AI Act: Introduces risk-based obligations for AI systems, with high-risk categories (e.g., AI in medical devices or machinery) requiring strict conformity assessment.
    Applicability: Adopted August 2024 (full text)

 

💡 Business implication: AI and data laws shift control of data flows away from pure ownership and toward regulated access and sharing. Companies must rethink business models that rely on exclusive data advantages and instead compete on quality, reliability, and responsible use of data.

 

Cybersecurity and Digital Resilience

 

  • Cyber Resilience Act (CRA): Embeds cybersecurity-by-design into connected products.
    Applicability: In force (full text)
  • NIS2 Directive: Expands cybersecurity obligations for a broad range of critical entities.
    Applicability: Adopted – Fully in force in 2027 (full text)
  • Digital Operational Resilience Act (DORA): Strengthens operational resilience in financial services, with spillover effects for suppliers.
    Applicability: In force (full text)
  • Cybersecurity Act: Provides EU-wide certification schemes for digital products.
    Applicability: In force (full text)

 

💡 Business implication: Cybersecurity regulation transforms digital security from an afterthought into a condition of market access. Products that fail to demonstrate resilience risk exclusion from supply chains, while early movers gain competitive advantage.

 

Product Safety

 

  • Product Liability Directive (revised): Extends liability to cover AI and digital features.
    Applicability: Adopted – Fully in force in 2026 (full text)
  • Radio Equipment Directive (RED): Ensures wireless and IoT devices meet safety and interoperability requirements.
    Applicability: In force (full text)
  • General Product Safety Regulation (GPSR): Overarching safety framework for all consumer products, including digital risks.
    Applicability: In force (full text)
  • Machinery Regulation (revised): Updates essential safety requirements, explicitly addressing AI and software safety.
    Applicability: Adopted – Fully in force in 2027 (full text)

 

❓ ❓ ❓ Confused about the difference between machines and products? A machine is an assembly with moving parts powered by energy (e.g., a washing machine, robot arm), while a product is the broader category of goods placed on the market (e.g., a smartphone). In short, all machines are products, but not all products are machines.

 

💡 Business implication: Product safety has expanded beyond mechanical hazards to include digital and AI risks. Companies must develop integrated safety strategies, treating hardware, software, and algorithms as a single system of responsibility.

 

 

Market Regulation

 

  • Digital Services Act (DSA): Enforces transparency and accountability rules for online platforms.
    Applicability: In force (full text)
  • Digital Markets Act (DMA): Addresses market dominance and ecosystem control by large digital "gatekeepers" (online platforms like app stores, search engines, and social networks that act as gateways between businesses and consumers).
    Applicability: In force (full text)

 

💡 Business implication: Compliance with DSA and DMA isn't just about legal alignment, it changes partner dynamics. Businesses will need to re-evaluate platform strategies, contracts, and marketing to stay discoverable in an increasingly regulated digital ecosystem.

 

 

This patchwork of obligations means AI compliance cannot be approached in isolation. Companies must embed AI within the broader regulatory framework — covering data governance, cybersecurity, liability, product safety, and sector-specific rules simultaneously.

 

How EU regulation works: The New Legislative Framework

The EU's system is based on the New Legislative Framework (NLF). This model separates what laws say from how compliance is achieved:

  • Legal acts define essential requirements: safety, transparency, robustness, accountability.
  • Directives set goals leaving it to each Member State to transpose them into national law.
  • Harmonised standards provide the technical details of how to meet these requirements. Once cited in the Official Journal of the EU, applying them creates a presumption of conformity.
  • Conformity assessment is mandatory before market access — either through internal control or, for higher-risk cases, third-party evaluation by a Notified Body.

This structure ensures laws remain technology-neutral and future-proof, while standards provide the evolving technical guidance companies need.

 

Challenges companies face

While the EU framework is comprehensive, many companies are not ready. Common challenges include:

  • Siloed approach: Focusing only on one law (e.g., the AI Act or GDPR) while ignoring overlaps.
  • Complexity: Multiple acts apply simultaneously, requiring cross-functional compliance strategies.
  • Documentation burden: New requirements for risk management, data governance, and transparency demand robust processes and for these to be documented in detail.
  • Capacity issues: Limited availability of interdisciplinary expertise may cause bottlenecks.

 

The risk is not just regulatory fines, but delays to market entry and loss of competitiveness.

 

Executive briefing: What senior leaders should ask themselves

For senior leaders, the question is not whether EU AI regulation will affect you, but how fast and how well you can adapt. Beyond the operational challenges, boards should be asking:

1. Do we know which of our AI use cases fall under high-risk categories?

2. Are we integrating AI requirements into our existing risk assessment and conformity checks?

3. Do we have visibility on emerging harmonised standards — and the resources to adopt them early? What is the value of moving faster than competitors?

 

The Role of Standards as Enablers

Standards are the practical bridge between regulation and implementation. For AI-embedded products, relevant standards include:

  • ISO/IEC 42001 – AI Management Systems
  • ISO/IEC 23894 – Guidance on Risk Management
  • ISO/IEC 12791 – Treatment of unwanted bias in classification and regression machine learning tasks
  • ISO/IEC 5259-X – Guidance on data quality
  • ISO/IEC DIS 27090 – Cybersecurity – guidance for addressing security threats to AI systems (currently under development)

 

By adopting these frameworks early, companies can operationalise compliance, streamline audits, and reduce the risk of costly redesigns. Browse through our overview of AI-related standards to view more.

 

Turning compliance into competitive advantage

Compliance is often seen as a burden, but in reality, it can be a competitive differentiator if achieved early in the game. Companies that align early with EU regulations and harmonised standards will:

  • Enter the EU market faster and with fewer obstacles,
  • Build trust with customers, regulators, and partners,
  • Reduce liability and long-term operational risk,
  • Position themselves as leaders in trustworthy, responsible AI.

 

📌 Lesson from the Field: The Cost of Overlooking CRA

An industrial equipment manufacturer focused only on the Machinery Regulation update and overlooked obligations under the Cyber Resilience Act (CRA). When they finally turned to cybersecurity compliance, CRA certification tests were fully booked ahead of the deadline as many companies rushed at once. Service costs tripled, and the firm faced significant delays and unplanned expenses before market launch.

The message is clear: companies that anticipate overlapping rules and move early turn compliance from a cost into a source of speed, savings, and market trust.

 

The take-home message

AI-embedded products face one of the most comprehensive regulatory landscapes in the world. 15 EU Regulations — spanning AI, data, cybersecurity, product safety, and markets — must be navigated in parallel. The New Legislative Framework shows the way: legal acts define the goals, harmonised standards provide the means, and conformity assessment unlocks the market.

For companies, the time to act is now. Below are three actions companies can take this year:

 

Next 12 Months – Leadership Priorities:

1. Map regulatory exposure across all AI-embedded products to identify high-risk use cases and overlapping obligations before they cause delays.

2. Appoint a cross-functional compliance lead to break silos between legal, engineering, product, and risk teams — turning compliance into a coordinated capability.

3. Pilot implementation against at least one AI standard to build internal know-how, test audit readiness, and reduce redesign risks.

Early movers won't just avoid compliance pitfalls — they'll shape internal capabilities, secure market access faster, and set the pace for trustworthy, competitive AI.

 

At Nemko Digital, we focus on helping product companies bridge the gap between AI innovation and regulatory compliance. With our deep expertise in AI governance, product conformity, and emerging standards, we support you in:

  • Understanding how the EU regulations affect your AI-embedded products,
  • Integrating AI-specific requirements into existing compliance systems,
  • Preparing for harmonised standards and future conformity assessments,
  • Building trust with regulators, customers, and partners.

 

If your products are evolving with AI, your compliance strategy must evolve too. Nemko Digital helps you turn regulation into a catalyst for responsible innovation — enabling safe, trusted AI in products.