AI Act published in the Official Journal of the EU
Today, 12 July 2024, the Artificial Intelligence Act (AIA) has been published in the Official Journal of the European Union. Explore it at this link. This is a moment of crucial importance for world legislation endeavours. We will have to wait until ...
Today, 12 July 2024, the Artificial Intelligence Act (AIA) has been published in the Official Journal of the European Union. Explore it at this link. This is a moment of crucial importance for world legislation endeavours. We will have to wait until 1 August, the date at which the Act is due to officially come into effect. From that day onward, the various compliance deadlines stipulated in the AIA will commence.
Adopted by the European Parliament and the Council, the Artificial Intelligence Act is the world’s first comprehensive framework for regulating AI across industrial and commercial sectors in the internal market. This landmark legislation addresses the use and deployment of AI systems, emphasising safety, transparency and accountability. The AIA aims to prevent the use of AI for repressive, manipulative, intrusive or discriminatory purposes. Citizens now have a clearer understanding of how their fundamental rights will be protected.
The AI Act has far-reaching implications in Europe and sets a global benchmark for regulating emerging technologies. The United States and other nations are closely monitoring these developments. The principles established by the AI Act are likely to influence similar regulatory frameworks in the U.S. The principles of the GDPR and the AI Act could inspire other regulatory advancements throughout the world, promoting the ethical implementation of AI across various legal landscapes.
The AI Act represents the EU’s progressive approach to innovation, balancing ethical regulation with the demands for safer and more responsible applications of artificial intelligence. If you are interested in how the AI Act can be implemented, you can learn more about etami, one of BDVA’s Task Forces. etami creates processes and tools to enable ethical, trustworthy and legal AI. etami translates European and global principles for ethical AI into actionable and measurable guidelines, tools and methods. Find out more!
etami’s “Fit for the AI Act”
etami just published “Fit for the AI Act”, a document meant to be a guide to help individuals and organisations better understand and prepare for the AI Act, which regulates AI systems based on their risk to individuals and society.
The AI Act revolves around two main principles: a risk-based approach and a horizontal application. The risk-based approach categorises AI systems by their potential harm to individuals or society (minimal, limited, high, or unacceptable) and imposes stricter regulations on higher-risk systems. For example, high-risk AI applications must adhere to stringent rules regarding transparency and data quality.
The horizontal application means the regulation spans across all sectors, ensuring uniform safety, transparency and accountability standards for all AI systems, regardless of their use case. The broad definition of AI in the regulation encompasses almost any software, affecting companies and research institutions alike, based on their interaction with AI.
For legal entities, the AI Act impacts any organisation involved in creating, using, or selling AI-related software or hardware. For individuals, the AI Act, along with the upcoming AI Liability Directive, provides a robust framework to protect rights and seek redress if harmed by AI.
Most provisions will come into effect in August 2026, with some exceptions like those concerning prohibited AI, which will be enforceable even sooner. Companies need to maintain an inventory of AI systems to ensure compliance, particularly avoiding prohibited uses like social scoring or real-time biometric identification.
High-risk AI systems have additional obligations, including risk management, data governance, technical documentation, transparency, human oversight accuracy, robustness, cybersecurity and conformity assessment. Compliance requires thorough preparation, as not all requirements are straightforward or fully defined yet.
Different parties (providers, suppliers, importers, deployers) must follow specific rules, ensuring AI systems meet the Act’s standards throughout their lifecycle. Users (deployers) are also liable for how AI systems are utilised.
Immediate actions include setting up a team or appointing dedicated personnel to understand and comply with the AI Act, creating an inventory of AI systems and collaborating with peers to share knowledge and best practices. Avoid rushing into audits, as standards are still under development. Start preparing now by understanding the AI Act and its implications for your operations. Seek expert advice where necessary and leverage collaborative networks for support. Explore the document now and find out more about etami & reach out to our task force here!