EU AI Act: navigating the future of AI regulation

October 8, 2024 | Thomas De Vos, AI Captain & Innovation Strategist at The Reference

Blog

As artificial intelligence continues to gain importance, the European Union is proactively implementing measures to ensure its ethical and responsible use.

One significant outcome is the EU AI Act, which came into effect on August 1, 2024. This pioneering legislation aims to regulate AI development and deployment throughout Europe.

Here at The Reference, we are not simply spectators; we are engaged contributors in this AI transformation. Let us explain what this means for businesses and how we can help you navigate these exciting yet complex changes.

The AI Act

A new digital era

The EU AI Act represents more than mere regulation—it embodies a broader vision. It is about adopting an environment in Europe where artificial intelligence is not only pioneering but is also secure and centered around human needs.

Officially taking effect in 2025, the Act aims to establish a digital environment that is both innovative and safe for European citizens.

For enterprises, this presents both a challenge and an opportunity to align the use of AI technology to gain a competitive edge, while committing to an ethical business approach towards their customers. 

Who is affected? 

Asking yourself, "Is this relevant to my business?" The answer is likely yes. The Act affects all entities—from large corporations to small startups. If you belong to any of these groups, you are impacted: 

  • Developers: If your company builds AI systems. 
  • Commercializers: If you bring AI products or services to market. 
  • Users: If you use AI in your daily operations. 
  • Importers and Distributors: If you are part of the supply chain for AI systems in the EU. 

Here at The Reference, we fit into multiple categories. Primarily, we are users—just like many of you—since we work with AI tools like OpenAI GPT. Additionally, we act as commercializers by offering AI-driven consultancy and services. Occasionally, we also serve as distributors when deploying our own- or third-party AI systems. We recognize that businesses may have concerns about adhering to these new regulations. Our goal is to simplify that process for you. 

The AI risk pyramid

From 'unacceptable' to 'minimal' 

The Act takes a risk-based approach to AI, categorizing systems from unacceptable to minimal risk: 

Unacceptable risk

Unacceptable risk AI systems are systems considered a threat to people and will be banned. They include: 

  • Cognitive behavioral manipulation of people or specific vulnerable groups: for example, voice-activated toys that encourage dangerous behavior in children.
  • Social scoring: classifying people based on behavior, socio-economic status, or personal characteristics.
  • Biometric identification and categorization of people.
  • Real-time and remote biometric identification systems, such as facial recognition.

 
Some exceptions may be allowed for law enforcement purposes. “Real-time” remote biometric identification systems will be allowed in a limited number of serious cases, while “post” remote biometric identification systems, where identification occurs after a significant delay, will be allowed to prosecute serious crimes and only after court approval. 

High risk

AI systems that negatively affect safety or fundamental rights will be considered high risk and will be divided into two categories: 

  • AI systems that are used in products falling under the EU’s product safety legislation. This includes toys, aviation, cars, medical devices, and lifts. 
  • AI systems falling into specific areas that will have to be registered in an EU database 
    • Critical infrastructures (e.g., transport), that could put the life and health of citizens at risk.
    • Educational or vocational training, that may determine the access to education and professional course of someone’s life (e.g., scoring of exams). 
    • Safety components of products (e.g., AI application in robot-assisted surgery).
    • Employment, management of workers and access to self-employment (e.g., CV-sorting software for recruitment procedures). 
    • Essential private and public services (e.g., credit scoring denying citizens opportunity to obtain a loan). 
    • Law enforcement that may interfere with people’s fundamental rights (e.g., evaluation of the reliability of evidence).
    • Migration, asylum, and border control management (e.g., automated examination of visa applications). 
    • Administration of justice and democratic processes (e.g., AI solutions to search for court rulings).

Limited risk

Limited risk is concerned with the lack of transparency in the use of AI. Therefore, it is important that AI systems that directly interact with people are developed to ensure that the person is informed that they are interacting with an AI system. 

Examples include chatbots and AI-generated content (text, images, videos, audio). Use these tools responsibly and with transparency. 

Minimal risk

Minimal or no-risk AI systems are those AI systems that do not fall into the three categories above. 

Most of today's AI applications, such as spell checkers, email classification, and text rewriting, fit into this category with only minimal regulation

Heavy penalties for non-compliance? Or an opportunity for responsible innovation?

Yes, there are fines for non-compliance, and they can be substantial— Penalties range from €7.5 million or 1.5% of turnover for minimal risk violations, up to €35 million or 7% of global turnover for unacceptable risk infractions.

The Act offers some flexibility to SMEs and startups by lowering the maximum fines. But instead of focusing on the penalties, think of this as an opportunity to lead the way in responsible AI innovation.

By aligning with the EU AI Act, businesses can build greater trust with customers and stakeholders, positioning themselves as leaders in a new, regulated AI era.

Two people showing each other something on a laptop screen

What this means for you 

Generative AI provides new unrivaled opportunities for marketers to create relevant personalized content at scale with great speed. At The Reference we embrace Generative AI to tackle content challenges or to provide superior user experience in digital challenges. In this context we mostly deal with limited and minimal risk systems.  

Here is what we do for limited risk AI: 

  • From the initial contact, let users know they are interacting with AI (unless it is obvious). 
  • Clearly label any AI-generated content. 
  • Make sure AI-generated text is disclosed in matters of public interest. 
  • Notify users and get their consent when they are exposed to systems that recognize emotions or categorize biometrics. 

For minimal risk systems, even though we are not legally required to, we hold ourselves to high ethical standards. We safeguard data privacy, prevent bias, and stay transparent about what our AI can and cannot do—because that’s simply good practice. 

For businesses deploying high-risk AI, there is a bit more oversight involved—like human oversight, performance monitoring, incident reporting, and ensuring compliance with provider guidelines.  

A person explaining something to his co-workers

What we can do for you

At The Reference, we are committed to embracing the changing AI landscape and using AI technology responsibly. We offer guidance in utilizing AI in ways that are compliant with EU regulations, ensuring transparency and adherence to legal requirements, particularly in customer-facing solutions.

Our services include advising on the smart use of AI to mitigate risks, connecting your ecosystem of internal data and tools to AI solutions via APIs and custom integrations, and setting up a controlled environment for safe internal use of AI by your employees. 

In addition, we help automate generative AI for content creationand maintenance, support customers in implementing composable commerce architectures, and transition into headless CMS solutions.

We understand that creating, managing, and updating content can be time-consuming and costly. Therefore, we aim to provide an end-to-end solution to tackle both the technical and content challenges of migrating to new platforms, significantly accelerating this transition with the use of generative AI. 

Our goal is to keep you ahead of new developments, ensuring that your AI strategies are innovative, compliant, and cutting-edge. We guide you through this transformation, allowing you to focus on innovation while we handle the complexities of compliance and integration

Two people sitting across from each other in a focused working booth

Looking ahead

The AI-powered future 

The EU AI Act sets a new standard for responsible AI. We see this as an exciting opportunity to build trust in AI technology and push innovation forward.

Whether you are just starting to explore AI or taking your existing AI tools to the next level, The Reference is here to help. Together, we can navigate this new era of regulated AI and make the most of the opportunities it presents. 

Let's continue the conversation


Did you like what you read? Are you eager to find out how we can help you further? Feel free to reach out, we would love to have a chat.

Let's talk

Keep reading with these insights

Unleashing AI to build a B2B commerce app from scratch

Discover how we leveraged generative AI and MACH components to build a fully operational B2B commerce app from scratch! See how AI tools like ChatGPT, Dall-E, and Midjourney were used to craft an immersive and personalized user experience. 

The future rewritten: understanding the impact of generative AI

Join our AI captain Thomas as he explores the rapid evolution of generative AI. Discover the opportunities, challenges, and the balance needed between technological progress and human values.

'Kan ik het ook'

Discover our feature in the VTM show 'Kan ik het ook' where industry experts, like ourselves, educate and inspire viewers about a variety of professions. In our episode, we challenged the host to create a striking campaign image with generative AI. While his efforts were impressive, it doesn't even compare to what our designers can accomplish with AI.