On November 26, 2023, the European Parliament and the European Council reached an agreement on the AI Act. Emphasizing law and not regulation!
The law, which was finally approved on December 20, 2023, and will enter into force on January 1, 2025, requires all developers and users of artificial intelligence in Europe to meet strict safety and human rights requirements.
The law establishes three types of artificial intelligence systems:
- AI systems with high risk - the ability to cause significant harm to humans, such as autonomous military or civilian systems for making decisions on humans. These will be approved by a regulatory authority before they are released to the market
- AI systems with medium risk - systems that can cause significant harm, but are not necessarily limited to humans. These will be required to meet strict safety requirements, such as reporting possible defects, protecting privacy and information security, and ensuring equal opportunities.
- Low-risk AI systems - systems that have no substantial risk of causing harm. These will be required to meet basic safety levels, such as privacy protection and information security.
There are additional categories of intelligence systems that the law refers to, such as:
- Education - systems used to assess students or make decisions about admission to school or university.
- Employment - systems used to make decisions about hiring, promoting or firing employees.
- Law enforcement - systems used to identify suspects or to make decisions about arrest or filing charges (buds already exist today in Alpha versions).
This is groundbreaking legislation in the field of artificial intelligence regulation. It is expected to affect Europe from the beginning of the year, Israel has determined that it will wait to see what other countries will do on the subject of regulation before it establishes its own regulation, but has begun by establishing a regulatory authority for AI, and has also issued a document of intent between the Ministry of Economy and the Ministry of Justice.
The law enacted by the European Parliament and the European Council obliges all EU countries, as well as EU member states and institutions, to comply with its requirements.
Other countries in the world may adopt similar regulations, based on the EU's AI law. For example, the United States, China, Japan, and, in my view, in the coming months or in the coming year also in Israel. What are the countries that have begun to develop their regulations in the field of AI.
Based on similar cases in the past, such as the Privacy Act, it is likely that many countries around the world will adopt regulations similar to the EU AI Act. These regulations may promote the safe and appropriate use of artificial intelligence.
The State of Israel has not yet officially announced whether it intends to adopt the European Union's artificial intelligence law for legislation within the country. However, Israel will likely adopt similar regulations, based on the progress of AI technology in the world.
Several factors may influence Israel's decision on this issue. One factor is Israel's desire to maintain international standards in the field of artificial intelligence. Another factor is Israel's desire to protect the human rights and privacy of its citizens.
In the end, the decision whether to adopt the EU's artificial intelligence law will be a political decision of the Israeli government.
The PDF document of the EU AI Act can be found on the European Commission website. The PDF contains the full text of the law, including all definitions, requirements, and exceptions.
