The European Union has officially published the new Artificial Intelligence Act (AI Act) in its Official Journal, with regulations set to take effect starting August 1st. Here’s what you need to know about Europe’s legislative move to redefine AI technology boundaries.
EU’s AI Act Enforces Security, Transparency, and Fundamental Rights
Primarily, the law aims to regulate AI technologies to enhance safety, transparency, and the protection of fundamental rights. From August 1st, AI technologies will no longer develop at random but will adhere to specific regulatory frameworks. What does this mean in practice?
The first phase of the law will commence within six months. During this period, certain prohibited AI applications, such as social credit systems and facial recognition databases, will be phased out. Additionally, real-time emotion recognition systems in schools and workplaces will be banned.
Nine months into the regulation, behavioral guidelines for AI developers will be introduced, prepared in collaboration with consulting firms by the EU AI Office. However, there are concerns about the potential influence of major corporations on these guidelines.
One year from now, new transparency requirements for AI models developed for general use will take effect. Tools like ChatGPT will need to demonstrate that their systems are safe and easily understandable by users. Ethical labeling regulations for generative AI and manipulated media are also on the horizon, aiming for greater transparency with deepfakes.
Companies training AI models will need to comply with copyright laws. If the model is not solely for research and development, permission will be required for text and data mining if the data belongs to others.
Timeline for Implementation of the AI Act:
- Definitions and prohibited applications will be effective from February 2, 2025.
- New notification requirements for AI models and database creation obligations will start from August 2, 2025.
- Classification of high-risk systems and related obligations will begin on August 2, 2027.
Violators of the law could face fines ranging from 1.5% to 7% of their annual global turnover. However, penalties for SMEs and startups will be more proportionate. EU countries will create regulatory sandbox environments to support SMEs and startups in developing and testing innovative AI systems under real conditions.
What are your thoughts on these regulations? How will they impact the development of AI technology? Share your views in the comments section below.
