🔆 The Yeah Act: On the Implementation of the EU AI Act

🗞️ Issue 14 // ⏱️ Read Time: 8 min

Hello 👋

The EU AI Act (the Act) has not even entered into effect, but there has been such high news coverage since the intense negotiations made the news in August 2023, that some of us feel it has been enforced at least 100 times. Even for those who follow these issues closely, it was difficult to keep up with the number of reactions, announcements, positions, negotiations, and news items that kept coming (and leaking!) out of the process.  

What is the EU AI Act?

The Act is a regulatory framework proposed by the European Union to govern the development and use of artificial intelligence (AI) within its member states. Its purpose is to support trustworthy AI in the EU and ensure that AI systems respect fundamental rights, safety, and ethical principles. The AI Act applies to providers and developers of AI systems marketed or used within the EU, regardless of whether those providers or developers are established in the EU or another country.

It categorises AI systems based on the level of risk they pose to health, safety, and fundamental rights, applying stricter regulations to higher-risk applications to ensure safety and compliance with ethical standards.

Timeline: EU AI Act Approval and Implementation 2024-2027

So, where are we now? 

We are a few months away from having a live AI Act in the EU. How is everybody feeling? Vibes vary. According to the Forum for the Kent A. Clark Center for Global Markets poll, opinions vary on the Act's potential impact on tech firms operating in the EU. 

Worth watching is how the EU as a political entity positions itself as the Act is implemented, as well as how other regions decide to regulate. Some want to see a Brussels effect, where the Act is used as an inspiration for other regions. 

64% of poll participants say the US needs similar regulation to impose testing requirements for powerful “foundational” models, prioritising safety over speed. Perhaps most importantly, any regulation on AI in other regions of the world should be locally developed and adapted to the contexts in which people and organisations operate.  

The Act, the world’s first comprehensive regulation of generative AI, is precisely that - it is the first, and will not be the last. There have been self-critical reflections from people deeply involved in the negotiations and legislative processes such as Dragos Tudorache and Kai Zenner. Keeping track of how the implementation goes in different regions will be vital for governments in the EU and organisations operating in the EU to see if the Act manages to fulfill its purpose - namely, supporting trustworthy AI in the EU and ensuring that AI systems respect fundamental rights, safety, and ethical principles. These are highly set and necessary goals and require a global dialogue and shared learning. Organisations focusing on human rights and digitalisation have raised concerns that standards for fundamental rights impact assessments are weak and that civic participation in the implementation and enforcement is not guaranteed in the Act. 

Don’t worry, prepare

Some investors are concerned about the implications of the Act and are considering moving their business elsewhere. For EU-based organisations, it remains important to stay competitive in the global market. Because of the unknown elements of the Act’s implementation, we see apprehension amongst organisations that use and develop AI technology. Alongside this, there is also a proactive energy and willingness to adapt and prepare for new technological developments and the regulation that comes with it. 

A passive, wait-and-see, attitude is not going to be beneficial for organisations in the long run. Even if several aspects of the Act are unclear, waiting for these clarifications will put organisations in a difficult position. 

Lumiera’s advice to organisations operating in the EU:

• Set an AI strategy and a roadmap: Align your organisation's short and long-term goals with EU AI Act requirements, taking the defined risk categories into consideration.

• Increase AI literacy and upskill the entire organisation: Enhance understanding of AI capabilities, limitations, and responsible usage.

• Implement robust data governance: Evaluate data sources, processing methods, and usage purposes, and establish data governance frameworks for compliance. 

• Commit to responsible AI practices. The base values of the Act can already be translated into guiding principles for development and usage. 

Until next time.
Emma, Sarah, and Allegra

Lumiera has gathered the brightest people from the technology and policy sectors to give you top-quality advice to navigate the new AI Era.

Follow the carefully curated Lumiera podcast playlist to stay informed and challenged on all things AI.