The Artificial Intelligence Act (AI Act) has been formally adopted, and the countdown to its entry into force has already begun. The AI Act sets general rules for high-risk, general-purpose AI systems (stand-alone AI systems). There are, however, some important nuances concerning such already existing AI systems.

Below, we present you a schedule that particularly providers of regulated AI systems should keep in mind when assessing the impact of the AI Act on their operators, including to existing AI systems and models. It should be remembered that most obligations in the AI Act are addressed to providers of these systems. These entities develop an artificial intelligence system or commission its development to place it on the market or put it into use under their trade name or trademark.

Key dates: General application of the AI Act

August 2, 2024: The AI Act enters into force after its publication in the Official Journal of the European Union (Article 113 of the AI Act).

August 2, 2026: On this date, numerous obligations arising from the Act enter into force, including those relating to high-risk AI Systems (Chapter III of the AI Act).

Nevertheless, certain obligations will be in force earlier and will be introduced gradually. These exceptions have the following deadlines:

February 2025: Six months after the publication the general provisions of the AI Act come into force (e.g. its scope, and definitions—Chapter I), as well as the provisions concerning prohibited AI systems (Chapter II, Article 5).

August 2025: Twelve months after the AI Act enters into force the following parts of it become effective:

• Provisions on stand-alone AI systems (general purpose) (Chapter V, Articles 50-56)
• Deadline for Member States to designate notifying authorities and state the notification procedure (Chapter III, Section 4, Articles 28-39)
• Penalties (Chapter XII, Articles 99 and 100)
• The duty of confidentiality applies to all legal and natural persons involved in the application of the AI Act (Article 78)

August 2027: Regulations on high-risk AI systems, covered by Article 6 (1) of the AI Act, i.e. AI systems that are security components or that are covered by the Union safety provisions listed in Annex I of the AI Act.

Existing high-risk AI systems and stand-alone (general-purpose) AI systems

Generally, the AI Act aims to prevent market disruptions and provide legal certainty for AI system operators and general-purpose system providers. The AI Act therefore provides exemptions from certain obligations or additional compliance time for operators of certain AI systems, such as stand-alone (general-purpose) systems, that were on the market before the entry into force of the AI Act. For such systems, the AI Act provides the following dates of entry into force of the obligations specified in it.

High-risk AI systems:
Operators (including suppliers, implementers, importers, and distributors) of high-risk AI systems that were placed on the market or put into service before August 2026 must only comply with the AI Act if those systems undergo “important changes” after this date.

Stand-alone AI systems (general purpose):
Suppliers of general-purpose AI systems that were placed on the market before August 2, 2025, have time until August 2, 2027, to adapt to the obligations arising from the AI Act.

For example, if a general-purpose system was placed on the market on January 1, 2025, which is within 12 months of the entry into force of the AI Act, the supplier has time until August 2, 2027, to comply with the AI Act regarding this system (less than 36 months).

However, if a provider plans to bring a general-purpose system to market on August 3, 2025, the provider must make it compliant with the applicable provisions of the AI Act.

What does it practically mean?

The first step is to analyze whether the AI Act will apply to the business model and activities of a given entity at all, even if such an entity uses various AI systems. Therefore, it is worth running such an analysis now to know whether to start preparing for the AI Act itself.

If the AI Act does apply, the above deadlines should be taken into consideration by product development, legal, and compliance teams, which should now start planning carefully and working closely to have all processes compliant with the AI Act requirements.

As a rule, however, the scope of the AI Act is quite narrow, so it is necessary to carefully assess whether the obligations arising from it apply to a given entity. This way, unnecessary complications related to compliance procedures that may not be necessary can be avoided.

All the AI system providers should prepare for the upcoming changes resulting from the Artificial Intelligence Act. Proper understanding and appropriate planning of activities are crucial to avoid unnecessary complications and ensure compliance with new regulations. It is worth starting to analyze the impact of the AI Act on one’s company’s operations now, to effectively adapt one’s processes and secure one’s position on the market. It should be remembered that proper preparation is not only a matter of compliance with the law but also a chance to build a competitive advantage in the dynamically developing artificial intelligence industry. The experts from Jabłoński Koźmiński law firm, who have extensive experience in the field of regulation and compliance, offer support in precisely adapting your company’s strategy to the requirements of the AI Act, helping to fully use the potential of artificial intelligence under the upcoming regulations.