EU AI Act Published on the Official Gazette and Will Come into Force on 01 August 2024
- The AI Act classifies AI according to its risk
- The General Purpose AI (GPAI), for example, Claude and ChatGPT, is specifically regulated
- Timeline of EU AI Act application
On 12 July 2024, the Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024 laying down harmonized rules on artificial intelligence (“EU AI Act”) has been published in the Official Gazette. After a four-year intense debate and scrutiny, this groundbreaking regulation is now becoming a reality. The EU AI Act will be enacted in 20 days following its publication (01.08.2024).
A brief overview of the EU AI Act's key provisions:
The AI Act classifies AI according to its risk
- Unacceptable risk AI systems are prohibited (e.g. social scoring systems, manipulative AI, and others).
- High-risk AI systems are regulated: AI systems shall comply with requirements and developers/deployers shall perform obligations (e.g. any AI systems in education, employment, healthcare, migration, etc. industry sectors).
- Limited-risk AI systems are subject to transparency obligations: developers and deployers must ensure that end-users are aware that they are interacting with AI (e.g. chatbots, deepfakes, etc.).
- Minimal-risk AI systems are unregulated (except for the obligation for AI literacy of employees).
The General Purpose AI (GPAI), for example, Claude and ChatGPT, is specifically regulated
- Providers of all GPAI models must establish technical documentation, and instructions for use as well as comply with the EU Copyright Directive, and publish a summary about the content used for training.
- Providers of GPAI models that present a systemic risk – open or closed – must also conduct model evaluations, and adversarial testing, track and report serious incidents and ensure cybersecurity protections.
Timeline of EU AI Act application
After entry into force, the AI Act will apply by the following deadlines:
- 6 months from publication: Prohibitions come into effect.
- 12 months from publication: Rules on GPAI models become applicable.
- 24 months from publication: Obligations for "stand-alone" High-Risk AI systems are enforced.
- 36 months from publication: Obligations for High-Risk AI systems embedded in products are enforced.
| Important: The EU AI Act will apply not only to AI providers and deployers located in the EU, but to any company developing, using, or distributing AI systems on the EU market or targeting EU consumers. |
Want to uncover the ins and outs of the EU AI Act and its implications for your business?
Our team at REVERA is here to guide you through the nuances of this revolutionary regulation. Get expert insights and practical advice on how to navigate the EU market's new AI landscape and avoid potential pitfalls. Reach out to us today to stay compliant and thrive in the new era of AI governance.
Authors: Kamal Tserakhau and Andrei Rutkouski
Dear journalists, the use of materials from REVERA website in publications is possible only after our written permission.
For approval of materials please contact e-mail: i.antonova@revera.legal or Telegram: https://t.me/PR_revera
