On 8th December, the European Union reached a provisional deal on the use of AI (artificial intelligence). These rules set by the European Union will regulate the AI use in all the countries under the EU and the act itself will become a landmark moment for Europe. This political agreement will set forward all the process in regard to the AI Act in the future.
The only place where government couldn’t find a mutual consensus was in the AI use for biometric surveillance and solutions on how to regulate AI tools like ChatGPT. But as the agreement was made the government might have overcome these points.
Now what are the key points of these agreements and how will this AI Act change the future of AI use in Europe.
Oliver Patel, who works at AstraZeneca as Enterprise Artificial Intelligence Governance Lead posted on his LinkedIn the image of the ‘AI Act cheat sheet.’ He quoted, “ Now the dust has settled on Friday’s announcement of a political agreement on the AI Act, it’s time to delve into the details,” with the image.
As AI Act becomes the first comprehensive AI law in the world which comes to force from 2026. The Act specifies on data safety, data sharing transparency and establishes a regulatory framework that will function concerning the market.
- In the AI Act the definition of AI has been given aligning with the recently updated AI definition by OECD. As per the new definition: ‘An AI system is a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment.”
- All dangerous AI systems will fall under the Prohibited AI, other categories are High-Risk AI, Limited Risk AI and then minimal Risk AI.
- Exemptions where AI Act will not be included are in defense sector, military, and national security along with R&D and partial exemption for open sources.
- Extraterritorial word has been used in regard to organizations outside the EU.
- A grace period of 6 to 24 months has been given.
- Organizations or people using High-risk AI will need to provide extensive requirements in regard to the AI use.
- In regard to generative AI, specific disclosure and transparency are needed.