01.08.2024 11:05:00
Дата публикации
The EU AI Act, a new law on artificial intelligence (AI), will come into force on August 1, 2024. This event will mark the beginning of a series of important milestones regarding compliance with the requirements set by the law.
The document, designed to regulate the use of relevant technologies, includes a ban on apps and services that threaten citizens' rights.
For example, biometric categorization based on sexuality, race or religion is prohibited, as well as the unauthorized extraction of facial images from the internet or surveillance cameras.
Transparency is a key aspect of the new legislation. Developers are required to provide summaries of the data used to train AI systems.
This requirement applies to all developers, including creators of "general AI systems" such as chatbots, which must respect copyright and demonstrate transparency regarding the use of their algorithms.
The law also sorts AI systems into risk categories: low, restrictive (chatbots and LLM), high (use by government agencies, police, etc.), unacceptable risk (rogue AI systems).
February 2, 2025 is the deadline for stopping the use of applications that may threaten the rights of citizens. Companies developing and using AI will have to stop using such technologies by this date.
From May 2, 2025, codes of practice containing detailed rules for companies will come into force. These documents will set standards for the legal use of AI in the jurisdiction of EU countries and outside it in certain cases.
From August 2025, companies working with “general AI systems” will be required to comply with new transparency requirements and respect copyright.
By August 2026, the rules will apply to all companies working with AI in the EU. Systems that fall into the “high risk” category, such as those used in infrastructure, healthcare and justice, have until August 2027 to fully comply with the new requirements.
Violations of the law will result in serious fines. Companies could face fines of up to €35 million, or 7% of their global annual revenue, for using prohibited systems. This underlines the seriousness of the EU’s approach to protecting citizens’ rights and regulating AI.
The full law can be found on the official website of the European Union.
(the text is translated automatically)
The document, designed to regulate the use of relevant technologies, includes a ban on apps and services that threaten citizens' rights.
For example, biometric categorization based on sexuality, race or religion is prohibited, as well as the unauthorized extraction of facial images from the internet or surveillance cameras.
Transparency is a key aspect of the new legislation. Developers are required to provide summaries of the data used to train AI systems.
This requirement applies to all developers, including creators of "general AI systems" such as chatbots, which must respect copyright and demonstrate transparency regarding the use of their algorithms.
The law also sorts AI systems into risk categories: low, restrictive (chatbots and LLM), high (use by government agencies, police, etc.), unacceptable risk (rogue AI systems).
February 2, 2025 is the deadline for stopping the use of applications that may threaten the rights of citizens. Companies developing and using AI will have to stop using such technologies by this date.
From May 2, 2025, codes of practice containing detailed rules for companies will come into force. These documents will set standards for the legal use of AI in the jurisdiction of EU countries and outside it in certain cases.
From August 2025, companies working with “general AI systems” will be required to comply with new transparency requirements and respect copyright.
By August 2026, the rules will apply to all companies working with AI in the EU. Systems that fall into the “high risk” category, such as those used in infrastructure, healthcare and justice, have until August 2027 to fully comply with the new requirements.
Violations of the law will result in serious fines. Companies could face fines of up to €35 million, or 7% of their global annual revenue, for using prohibited systems. This underlines the seriousness of the EU’s approach to protecting citizens’ rights and regulating AI.
The full law can be found on the official website of the European Union.
(the text is translated automatically)
