The European Union’s AI Act, the world’s first comprehensive regulation for artificial intelligence, officially starts on 1 August 2024, aiming to balance innovation and risk management.
The European Union’s AI Act, a comprehensive regulation aimed at tackling the potential risks associated with artificial intelligence, officially came into effect on 1 August 2024. This legislation, the first of its kind globally, reflects the EU’s effort to create a framework that balances innovation and regulation by implementing a risk-based approach to AI applications and technologies.
The initiative to develop the AI Act began in April 2021 when the European Commission proposed a regulatory framework to ensure AI technologies remain “human-centred.” The proposal aimed to foster trust among EU citizens while providing businesses with clear guidelines on AI deployment. Since then, the regulation has undergone intense scrutiny and discussions, leading to the final political agreement in December 2023 and adoption in May 2024.
The AI Act categorises AI applications based on the level of risk they pose. “Unacceptable risk” applications, including those that utilise harmful subliminal techniques or social scoring, are banned, albeit with various exceptions. Conversely, “high-risk” applications, such as those used in critical infrastructure, law enforcement, education, and healthcare, must undergo conformity assessments before and after market deployment. These assessments ensure compliance with EU standards on data quality, transparency, cybersecurity, and human oversight, among other criteria. Public bodies deploying high-risk systems are required to register these systems in an EU database.
A “medium-risk” category requires transparency measures, informing users they are interacting with AI-generated content, applicable to technologies such as chatbots. Recently, the regulation has been adjusted to address emerging concerns around General Purpose AI (GPAI), driven by the proliferation of generative AI tools like ChatGPT. GPAIs, typically underpinning various AI applications, are subject to transparency rules and risk assessments for those considered to pose systemic risks, measured by computing power used during model training.
As Europe leads in setting worldwide standards for AI regulation, the Act aims to build trust within the local AI ecosystem while managing the inherent risks of AI technologies. This regulatory approach could potentially influence global AI standards, encouraging a conversation around ethical AI development.
The implementation of the AI Act features a staggered approach across timelines extending into 2027. Key compliance deadlines span from 2025, when prohibited use case rules begin to apply, to comprehensive governance and transparency obligations by 2027. This phased approach aims to provide businesses ample time to comply while allowing regulators to define detailed compliance requirements in this rapidly evolving technological landscape.
Oversight and enforcement are split between an EU-centralised AI Office and member state-level authorities. Penalties for non-compliance are substantial, with fines reaching up to 7% of global turnover for serious breaches. How effective this multi-level enforcement structure operates remains to be seen as regional authorities begin their assessment and compliance investigations.
While the AI Act represents a significant regulatory milestone, it is clear that the pace of AI development will require ongoing adjustments to these rules, ensuring they remain relevant and effective in managing the technology’s risks. The EU continues to develop comprehensive guidance and standards essential for the law’s successful application, reinforcing its role as a pioneer in AI regulation on the global stage.
Source: Noah Wire Services
- https://www.whitecase.com/insight-alert/long-awaited-eu-ai-act-becomes-law-after-publication-eus-official-journal – Corroborates the EU AI Act’s entry into force on 1 August 2024 and its comprehensive risk-based approach to regulating AI systems.
- https://iapp.org/resources/article/eu-ai-act-timeline/ – Provides details on the timeline for the implementation of the EU AI Act, including key dates and milestones.
- https://www.ropesgray.com/en/insights/viewpoints/102jd2c/eu-ai-act-published-in-the-official-journal-of-the-european-union-clock-starts-f – Explains the entry into force of the EU AI Act on 1 August 2024 and the staggered implementation timeline.
- https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai – Describes the EU’s approach to AI regulation, including the establishment of the European AI Office and the aim to foster trustworthy AI.
- https://insightplus.bakermckenzie.com/bm/data-technology/european-union-eu-ai-act-published-dates-for-action – Details the ban on ‘prohibited systems’ and the phased implementation approach of the EU AI Act.
- https://www.whitecase.com/insight-alert/long-awaited-eu-ai-act-becomes-law-after-publication-eus-official-journal – Corroborates the categorization of AI applications based on risk levels and the requirements for high-risk applications, including conformity assessments.
- https://iapp.org/resources/article/eu-ai-act-timeline/ – Provides information on the transparency measures for medium-risk AI applications and the adjustments for General Purpose AI (GPAI).
- https://www.ropesgray.com/en/insights/viewpoints/102jd2c/eu-ai-act-published-in-the-official-journal-of-the-european-union-clock-starts-f – Explains the oversight and enforcement structure, including the role of the EU-centralised AI Office and member state-level authorities.
- https://digital-strategy.ec.europa.eu/en/policies/regulatory-framework-ai – Describes the penalties for non-compliance, including fines up to 7% of global turnover for serious breaches.
- https://www.whitecase.com/insight-alert/long-awaited-eu-ai-act-becomes-law-after-publication-eus-official-journal – Highlights the need for ongoing adjustments to the AI Act to ensure it remains relevant and effective in managing AI risks.
- https://iapp.org/resources/article/eu-ai-act-timeline/ – Details the EU’s ongoing development of comprehensive guidance and standards for the successful application of the AI Act.
