[go: nahoru, domu]

Artificial Intelligence Act: Difference between revisions

Content deleted Content added
Refactored the leading paragraphs.
mNo edit summary
Line 33:
It covers all types of AI across a broad range of sectors, with exceptions for AI systems used solely for military, national security, research and non-professional purposes.<ref>{{Cite news |date=9 December 2023 |title=Artificial intelligence act: Council and Parliament strike a deal on the first rules for AI in the world |url=https://www.consilium.europa.eu/en/press/press-releases/2023/12/09/artificial-intelligence-act-council-and-parliament-strike-a-deal-on-the-first-worldwide-rules-for-ai/ |access-date=January 6, 2024 |work=Council of the EU |archive-date=10 January 2024 |archive-url=https://web.archive.org/web/20240110092304/https://www.consilium.europa.eu/en/press/press-releases/2023/12/09/artificial-intelligence-act-council-and-parliament-strike-a-deal-on-the-first-worldwide-rules-for-ai/ |url-status=live }}</ref> As a piece of product regulation, it does not confer rights on individuals, but regulates the providers of AI systems and entities using AI in a professional context.<ref name=":4">{{Cite web |last=Mueller |first=Benjamin |date=2021-05-04 |title=The Artificial Intelligence Act: A Quick Explainer |url=https://datainnovation.org/2021/05/the-artificial-intelligence-act-a-quick-explainer/ |access-date=2024-01-06 |website=Center for Data Innovation |language=en-US |archive-date=14 October 2022 |archive-url=https://web.archive.org/web/20221014172842/http://datainnovation.org/2021/05/the-artificial-intelligence-act-a-quick-explainer/ |url-status=live }}</ref>
 
The Act classifies [[#Exemptions|non-exempted]] AI applications by their risk of causing harm. There are four levels –&nbsp; unacceptable, high, limited, minimal – plus an additional category for general-purpose AI. Applications with unacceptable risks are banned. High-risk applications must comply with security, [[transparency (behavior)|transparency]] and quality obligations, and undergo [[Conformity assessment|conformity assessments]]. Limited-risk applications only have transparency obligations, while minimal-risk applications are not regulated. For general-purpose AI, transparency requirements are imposed, with additional evaluations for high-capability models.<ref name=":1">{{Cite web |last=Espinoza |first=Javier |date=December 9, 2023 |title=EU agrees landmark rules on artificial intelligence |url=https://www.ft.com/content/d5bec462-d948-4437-aab1-e6505031a303 |url-status=live |archive-url=https://web.archive.org/web/20231229130308/https://www.ft.com/content/d5bec462-d948-4437-aab1-e6505031a303 |archive-date=29 December 2023 |access-date=2024-01-06 |website=Financial Times}}</ref><ref name=":2">{{Cite news |title=EU AI Act: first regulation on artificial intelligence |url=https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence |access-date=2024-01-06 |work=European Parliament News |language=en |archive-date=10 January 2024 |archive-url=https://web.archive.org/web/20240110162551/https://www.europarl.europa.eu/news/en/headlines/society/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence |url-status=live }}</ref>
 
The Act also creates a European Artificial Intelligence Board to promote national cooperation and ensure compliance with the regulation.<ref>{{Cite web|last1=MacCarthy|first1=Mark|last2=Propp |first2=Kenneth |date=2021-05-04 |title=Machines learn that Brussels writes the rules: The EU's new AI regulation |url=https://www.brookings.edu/blog/techtank/2021/05/04/machines-learn-that-brussels-writes-the-rules-the-eus-new-ai-regulation/ |url-status=live |archive-url=https://web.archive.org/web/20221027115538/https://www.brookings.edu/blog/techtank/2021/05/04/machines-learn-that-brussels-writes-the-rules-the-eus-new-ai-regulation/ |archive-date=27 October 2022 |access-date=2021-09-07 |website=Brookings |language=en-US}}</ref> Like the EU's [[General Data Protection Regulation]], the Act can apply [[Extraterritoriality|extraterritorially]] to providers from outside the EU if they have users within the EU.<ref name=":4" />