Spain – A State Agency for the Supervision of Artificial Intelligence and Control of Algorithms
Spain has approved a law for the establishment of an agency for the supervision of AI and the control of algorithms (AESIA). The government has allocated 5 million euros for this purpose in the budget. The agency will be fully operational by the end of 2023. The new body will conduct oversight on AI systems and algorithms used by both government agencies and private enterprises.
What: Policy-oriented document
Impact Score: 2
Target audience: Policy makers, sectorial organisations, companies
Key takeaways for Flanders: For policymakers in Flanders and Belgium, it will be interesting to see how Spain will be the first EU member state to develop supervision of AI through a government body. However, it is essential that the new body's remit does not overlap or risk conflicts with already existing government bodies. In Spain, for example, one has to consider the Spanish Data Protection Agency, the National Commission for Competition or the Spanish cybersecurity agency.
Summary
The state agency for AI supervision and algorithm control will have legal personality and will operate independently from other government services. This setup is intended to enable the body to act objectively, transparently, and impartially. The agency will closely monitor Spain's national AI strategy and coordinate and promote actions to oversee government institutions and private organizations for the lawful use of AI systems and algorithms. It must ensure that this technology complies with existing legal regulations and ethical requirements for algorithms, including preventing algorithms from reproducing dynamics of discrimination based on gender, race, or sexual orientation.
It will also be responsible for creating regulated testing environments that enable the responsible launch of innovative high-risk or general-purpose AI systems. Furthermore, the agency must oversee compliance with the "Rider law" (legislation protecting platform workers) and have the authority to impose sanctions. The specific details of its responsibilities will be announced later this year.
With the establishment of the agency, Spain is preparing for the EU's AI Regulation. Spain has also set up the first regulatory sandbox for AI regulation, outlining various responsibilities proportional to the risk level of a specific AI system. This sandbox facilitates the testing and experimentation of AI applications within a controlled environment.
AI Code of Conduct
Currently, policymakers from the EU and the US are collaborating on a Code of Conduct for AI. The aim is to establish standards for the use of AI technology, serving as an interim measure until comprehensive legislation, such as the AI Regulation, is adopted.
AESIA will cooperate with the EU to support the development of this code of conduct.