The EU AI Act Is In Force From Today (1 August 2024)

The new EU Regulation No. 2024/1689 on artificial intelligence - the EU AI Act - has finally been adopted and will now enter into force today, 1 August 2024, and the obligations under the Act will begin to apply incrementally (see below). 

The new regulatory framework deals with the regulation of AI systems and their uses and emphasises the need to balance technological advancement with the protection of intellectual property ("IP") rights and transparency.

Here are the key IP points:

1. Compliance with Copyright Law: Providers of general-purpose AI models (capable of various applications) must ensure adherence to EU copyright and related rights. The text and data mining (TDM) exception must be carefully applied, especially with regard to content protected by opt-out rights, to respect the rights of content creators. Such content is often protected by copyright and requires permission from rights holders.
2. Copyright Policy: Providers must implement internal policies to safeguard third-party IP rights, preventing unauthorized use during AI training and ensuring compliance with opt-outs from TDM.
3. Transparency Reporting: AI providers are required to maintain and update detailed technical documentation about their systems, including training processes, test results, and other essential information for model development.
4. Monitoring by the AI Office: The AI Office will oversee compliance with copyright obligations and the publication of required content reports.
5. Information Obligations: Users must be informed when interacting with AI systems to promote transparency and trust.

What are the most significant deadlines to be aware of?

➡ Early 2025 - Businesses will be prohibited from using any AI systems carrying unacceptable risk

➡ May 2025 -  The EU AI Office is to make available codes of practice for general purpose AI models (including foundation models, which train on large amounts of data)

➡ August 2025 - Businesses will have to abide by the rules on general purpose AI models

➡ August 2026 - Most of the remaining obligations, including for the specific standalone high-risk AI categories and the obligations for the limited risk AI categories 

➡ August 2027 - Obligations in relation to high-risk AI systems that are products or safety components in products already subject to third-party conformity assessment requirements under sectoral EU regulation

In every sector, businesses will thus have to adapt quickly to the new regulatory framework, non-compliant companies will be exposed to the risk of fines of up to 7% of annual turnover for prohibited AI systems.

For more on the EU AI Act's obligations and deadlines, see HSF's post here.