Member states have 12 months to nominate national competent authorities tasked with compliance.
ADVERTISEMENTEurope’s AI Act, the world's first attempt to regulate AI systems according to a risk-based approach, will get the final nod in the European Parliament tomorrow (13 March), paving the way for the rules to finally enter into force. However, when it comes to oversight; member states are still in the early stages of determining which regulator is best placed to oversee compliance.
Tomorrow, the lawmakers will vote on the text without the minor linguistic changes made by the lawyers during the translation phase. These also need to be formally approved, either by a separate vote in the April plenary or by a formal announcement. The rules will then be published in the EU official journal which is likely to happen in May.
Under the AI Act, machine learning systems will be divided into four main categories according to the potential risk they pose to society. The systems that are considered high risk will be subject to stringent rules that will apply before they enter the EU market.
In November, bans on prohibited practises specified in the AI Act will apply. The general-purpose AI rules will apply one year after entry into force, May 2025, and the obligations for high-risk systems in three years. They will be under the oversight of national authorities, supported by the AI office inside the European Commission.
A European Commission spokesperson told Euronews that the countries have 12 months to nominate relevant national competent authorities, and the commission “awaits notification of these appointments in due course”.
Recruitment
Spain was the first EU country to set up an Agency for the Supervision of Artificial Intelligence (AESIA) in A Coruña in 2023. In other countries such as the Netherlands, the data protection authority set up a department dealing with algorithms last year. The office currently has 12 employees and the expectation is that it will grow to 20 people this year.
In the case of Ireland, the Department of Enterprise, Trade and Employment will lead the development of the national implementation plan for the AI Act. A spokesperson told Euronews however that “as the EU legislative process has not yet completed, and the Act has not yet been adopted, it would be premature to speculate about the national arrangements for enforcement.”
The Luxembourg Department of Media, Telecommunications and Digital Policy said that the country is "working hard to implement the AI Act".
"We are consulting with all the relevant stakeholders, first and foremost the existing authorities and regulators that will have a role to play in the new governance framework. An important aspect for us is an efficient coordination between regulators, as we want to make it as clear as possible for businesses and citizens to interact with the new legislation," a spokesperson said.
Meanwhile, the commission also began its recruitment process for policy and technical jobs at the AI Office, which will help with delivering a harmonised approach of the rules across the member states via information exchanges. The deadline for applications is 27 March.
Trade organisations have warned about the lack of implementation and enforcement now that the rules are on the brink of entering into force. CCIA Europe, which represents Big Tech companies, previously warned that the implementation will be crucial “to not overburden companies” that try to innovate.
Digital Europe, representing both tech companies and national trade associations, called in the AI Act trilogue negotiations for a 48-month transitional period to ensure the whole ecosystem is ready, as well as for the timely availability of harmonised standards for industry to comply with the rules.
