Alberto Gago at the Helm of Aesia: Pioneering AI Regulation in Spain
Since December, Alberto Gago, a 39-year-old from Zaragoza, has taken on the role of director of the Spanish Agency for the Supervision of Artificial Intelligence (Aesia). This agency is the first of its kind within the European Union and is responsible for ensuring that artificial intelligence (AI) systems in Spain adhere to legal standards. Gago has a strong background in this area, having served as an advisor to Carme Artigas, the former Secretary of State for Digitalization and AI, during the extensive negotiations leading up to the European AI Regulation, which Spain facilitated during its presidency of the EU.
Understanding the European AI Regulation
The European AI Regulation categorizes systems based on the risk they pose to society. It classifies them into four levels: minimal risk (like spam filters), limited risk (including generative AI), high risk (involving critical infrastructures, healthcare, education, and justice), and unacceptable risk (which includes systems such as social scoring and real-time biometric identification—these are strictly banned).
Each EU member state is tasked with ensuring that prohibited systems are not utilized and verifying that high-risk systems meet the necessary legal controls within the upcoming months. Aesia is the organization designated to handle these responsibilities in Spain. Based in La Coruña, Aesia currently employs 35 people, with plans to expand to 50 by the end of the year and continue growing in subsequent years. Gago emphasizes the need for a diverse team, stating, “We have computer engineers, economists, sociologists, philologists, and jurists. All are essential to understand the societal implications of AI.”
Aesia's Role and Responsibilities
Q: The Aesia is relatively unknown to the public. What are its specific powers?
A: Aesia is still in its developmental phase but is already operating at full capacity. Our primary objective is to build trust in AI usage while providing legal certainty for companies. Right now, we are focusing on monitoring AI systems as mandated by the AI Regulation. Currently, the regulations concerning prohibited systems are in effect; we have examined thousands of systems and confirmed that no banned AI systems are currently available in the Spanish market.
Q: How do you monitor compliance?
A: We use various channels for monitoring. Complaints can be submitted via a dedicated mailbox. We actively monitor media and market reports and occasionally consult suppliers regarding specific systems. Additionally, we employ a systematic approach to scan different sectors, prioritizing areas like critical infrastructure and essential public and private services. However, we're still in the process of negotiating the Digital Omnibus Law, which will define how we evaluate high-risk systems; we expect clarity on this matter soon.
Training and Supporting AI Governance
Q: What additional powers does Aesia possess?
A: Aesia is also responsible for training and educating the public about AI. In our first year, we have responded to around 700 inquiries regarding regulation applications and the associated risks and benefits of AI. Our mandate includes boosting innovation by publishing guides and testing AI systems in practice. On the international front, we aim to contribute to the development of institutional AI governance frameworks, aligning with the European Commission and collaborating with institutions like the Valencia AI governance laboratory. We also plan to establish an ideas laboratory to study the societal and economic impacts of AI.
Q: When will this think tank be launched?
A: We aim to set it up by the end of March. This initiative will gather insights from prominent national experts to explore social and economic challenges posed by AI, such as misinformation and labor market effects. We encourage engagement with the community, including schools and universities, to better understand real-world concerns.
Future Regulation and AI Challenges
Q: You mentioned the Omnibus Law, which will amend the AI regulation shortly after its implementation. Do you believe this is necessary?
A: No regulation is flawless. What's crucial, as endorsed in the EU, is that we don't devolve into deregulation but rather focus on refining our approach to better manage these complexities.
Q: Recently, the President of the Government requested the Prosecutor's Office to investigate social media platforms for disseminating inappropriate content using AI. Has Aesia provided any technical support?
A: Spain has a history of advocating for rights. The President's decision to prohibit minors from accessing inappropriate content aligns with practices in countries like Australia and France. This is about ensuring that rights in the physical realm are upheld in the digital landscape. If companies cannot guarantee respect for these rights, there must be consequences. Aesia will support these efforts within its jurisdiction, including educating youth about AI and assisting in investigations if requested.
Q: Has Aesia already participated in such investigations?
A: We are awaiting requests for assistance, but so far, none have been made.
