EU AI Act - key take-aways!
Last Tuesday, May 21, the 27 member countries of the European Union gave their final approval to ground-breaking legislation to regulate Artificial Intelligence systems! 🇪🇺
This AI Act is a European regulation on AI aiming at guaranteeing that all AI systems and models commercialized within the EU are used ethically, safely and in compliance with EU fundamental rights. It is the first comprehensive regulation on AI by a major regulator anywhere in the world.
Claire, Ophélie and Louis dived in and made a quick recap for you 🔎
How’s the European approach different?
🇺🇸 US – More fragmented vision – federal acts, regulation fragmented by sectors. US Executive Order on AI: guidelines, requiring federal agencies to only buy and use responsible solutions. The reason being that this technology still needs room to grow and develop before broad, binding regulation becomes necessary.
🇨🇳 China – Highly iterative approach. China’s fundamental approach to AI governance remains focused on mitigating harm to individuals and maintaining social stability and state control, while aiming for global AI leadership and influencing the debate on AI regulation.
🇬🇧 UK – No legislation yet. First summit on AI safety.
🇪🇺 Europe – Potential Brussels Effect – companies will build their practices to comply with the EU AI Act, which should enable them to comply de facto with most US/Chinese requirements. More prescriptive and risk based approach.
What are the main principles of the risk based approach?
Limited risk – Applications not explicitly banned or listed as high-risk are largely left unregulated.
High risk – High-risk applications, such as a CV-scanning tool that ranks job applicants, are subject to specific legal requirements.
Unacceptable risk – Applications and systems that create an unacceptable risk, such as government-run social scoring of the type used in China, are banned and may not be commercialized within the EU.
Wanna know more about who is concerned and how will it shape the market? Click here.
This AI Act is a European regulation on AI aiming at guaranteeing that all AI systems and models commercialized within the EU are used ethically, safely and in compliance with EU fundamental rights. It is the first comprehensive regulation on AI by a major regulator anywhere in the world.
Claire, Ophélie and Louis dived in and made a quick recap for you 🔎
How’s the European approach different?
🇺🇸 US – More fragmented vision – federal acts, regulation fragmented by sectors. US Executive Order on AI: guidelines, requiring federal agencies to only buy and use responsible solutions. The reason being that this technology still needs room to grow and develop before broad, binding regulation becomes necessary.
🇨🇳 China – Highly iterative approach. China’s fundamental approach to AI governance remains focused on mitigating harm to individuals and maintaining social stability and state control, while aiming for global AI leadership and influencing the debate on AI regulation.
🇬🇧 UK – No legislation yet. First summit on AI safety.
🇪🇺 Europe – Potential Brussels Effect – companies will build their practices to comply with the EU AI Act, which should enable them to comply de facto with most US/Chinese requirements. More prescriptive and risk based approach.
What are the main principles of the risk based approach?
Limited risk – Applications not explicitly banned or listed as high-risk are largely left unregulated.
High risk – High-risk applications, such as a CV-scanning tool that ranks job applicants, are subject to specific legal requirements.
Unacceptable risk – Applications and systems that create an unacceptable risk, such as government-run social scoring of the type used in China, are banned and may not be commercialized within the EU.
Wanna know more about who is concerned and how will it shape the market? Click here.
Da yours,
The daphni team 👾
