AI Regulation Around the World: Spain

With its proposed harmonised rules on artificial intelligence, known as the EU AI Act, the European Commission is aiming to create greater trust around the use of AI by regulating the AI systems available in the EU Market. Although the act could arguably become the global gold standard for AI regulation, EU member states, such as the Netherlands and Spain, are starting to propose their own laws, demonstrating their commitment to regulating AI.

Despite not contributing to the development of the AI Act in the form of a presidency compromise text, Spain has played an active role in supporting the trial enforcement of the obligations of the Act by establishing the first regulatory sandbox to allow experiments in a controlled environment. Further demonstrating its commitment to governing AI to reduce the harms that it can pose, Spain has published a National AI Strategy and announced the launch of Europe’s First AI Supervisory Agency to oversee the implementation of the Strategy and AI Act and has targeted platform-based worker rights with its so-called Rider Law. In this blog post, we provide a high-level overview of some of the AI-relevant initiatives introduced by the Spanish Government in recent years.

Key takeaways:

• Spain has launched the first regulatory sandbox for the EU AI act to create a controlled environment for experimenting with the obligations of the Act in a controlled environment
• Outside of its contributions to the EU AI Act, Spain has also published a National AI strategy to ensure that AI does not negatively impact values, rights, or social welfare and instead reduces social inequality and promotes innovation
• To support the implementation of the Strategy and EU AI Act, Spain has also announced that it will launch Europe’s first AI Supervisory Agency to monitor and mitigate AI risks
• Taking a more targeted approach, Royal Decree-Law 09/21 amends the Workers Statute to give employment rights to platform-based delivery riders and transparency rights to all platform workers

Spain’s regulatory sandbox for the EU AI Act

As part of the EU AI Act, EU member states are encouraged to launch sandboxes to create a controlled environment for experiments and testing of AI systems against the requirements of the EU AI Act under the supervision of relevant authorities. Spain is leading these efforts, having launched the first sandbox aimed at testing the requirements of the legislation, as well as how conformity assessments and post-market activities may be overseen.

With a budget of 4.3 million euros, the pilot will be financed by the Spanish government’s recovery and resilience funds as part of the Spanish National AI strategy. The pilot was expected to begin in October 2022, with the results being published by the end of 2023. Deliverables from this trial include documentation of obligations and how they can be implemented, and methods for controlling and following-up that can be applied by those supervising national authorities responsible for implementing the regulation.

Small and medium sized enterprises will be given priority access to the sandbox to support the reduction of barriers that these companies might face when launching their AI systems under the new regulation. However, it is important to note that while this is a controlled environment and provides an opportunity to test systems, those that participate in the sandboxes are still liable for any harm inflicted on third parties that results from their activities.

Spain’s National AI Strategy

In addition to the work Spain is doing to support the EU AI Act, the country has also launched its own initiatives to govern how AI is being used in the country. Indeed, published in November 2020, Spain’s National AI Strategy signals the importance the Spanish government is placing on ensuring that AI does not negatively impact values, rights, or social welfare, and instead reduces social inequality and promotes innovation. Part of the Digital Spain 2025 Agenda launched in July 2020 – which aims to promote the digital transformation of Spain over the next five years through almost 50 measures – the National AI strategy is a flexible framework that will shape Spain’s approach to AI, with a focus on the wellbeing of citizens.

Backed by a public investment of €600 million, the strategy is defined by six axes:

Within each of these axes is a series of actionable measures that will support the realisation of the strategy, including investment in (post) doctoral programs for AI research, the launch of a SpAIn talent hub, creation of the Data Office and Chief Data Officer, development of a National Green Algorithms Program, and the creation of a Digital Bill of Rights. Through these measures, the strategy aims to achieve seven key objectives:

1. Scientific excellence and innovation in AI

2. Projection of the Spanish language without worldwide development of AI tools, technologies, and applications

3. Creation of qualified employment and the promotion of training and education in AI

4. Transformation of productivity for sustainable and inclusive economic growth

5. Generate an environment of trust surrounding AI

6. Promote humanist values in AI, with a focus on ensuring societal wellbeing and developing an ethical framework to protect individual and collective rights

7. Promote inclusive and sustainable AI to reduce the gender gap and digital divide

Spain’s AI Supervisory Agency

To support the implementation of the National AI Strategy, Spain has announced that it will create Europe’s first Supervisory Agency for AI to monitor and mitigate AI risks. Predicted to be launched in late 2023, the Agency will not only be responsible for the supervision and monitoring of projects carried out as part of the National AI Strategy but will also support the enforcement of the EU AI Act when it comes into effect. Indeed, AESIA (Agencia Española de Supervisión de la Inteligencia Artificial) is likely to be designated as Spain's national competent authority for supervising the application and implementation of the Act, as required by Recital 77.

To achieve this, the Agency would be given €5 million at its disposal, a figure that is intended to be increased before the launch of the agency if budgetary restrictions permit. Much of these funds will go towards employing a 40-strong staff with expertise in technology, administration, and law. Social scientists may also be hired to join this team to research the social impact of AI.

Royal Decree 09/21 - The Rider Law

As well as these broader efforts to regulate AI, Spain has also begun to take a more targeted approach, seeking to regulate particular applications. Known as the rider law, Spain’s Royal Decree-Law 9/2021 (RDL 9/2021) came into effect on 12 August 2021 and targets delivery workers whose work is coordinated through digital platforms. Amending the Workers' Statute to give delivery riders a presumption of employment, the Royal Decree-Law entitles them to the same protections and benefits as other employees. This reflects similar action that has been taken in the UK, France, and Italy, where previously self-employed delivery riders were recognised as employees following court judgements, giving them access to minimum pay and pensions.  However, since this is only a presumption of employment, employers could provide evidence to the contrary that indicates that they do not exercise their powers of organisation, direction and control over platform-based delivery workers. Nevertheless, this provision is essential to ensure safer and fairer working conditions for delivery workers with flexible contracts.

In addition to the rights given to delivery riders specifically, the Royal Decree-Law also sets out algorithmic transparency requirements concerning all workers who work through a digital platform. Modifying Article 64.4 of the Workers Statute, employers must now inform employees’ legal representatives or works council of the parameters, rules and instructions on which the algorithms or AI systems are based. This, therefore, gives platform workers greater transparency about how decisions are made about them and the factors that can influence these decisions, and could make it easier to bring about complaints and legal action in the event that workers feel that they are treated unfairly by the platforms.

To support Spain’s efforts towards greater transparency and accountability for algorithmic systems, the Ministry of Labor has published guidelines on algorithmic data in the workplace to bring obligations around algorithmic systems in a labour context together and provide a tool to specify and systematize information obligations. The guidelines define algorithms and outline how automated decision systems can be used in the workplace, including hiring decisions, monitoring and surveillance, and work management, and summarises how GDPR applies to algorithmic systems, as well as how the transparency obligations required by RDL 9/2021 can be met. Divided into four sections, the tool covers information that should be disclosed regarding:

1. General information on the use of algorithms or AI to make automated decisions – the decisions the system is used for, the technology used by the algorithm, the software used and who supplies it, and the use of human interventions.

2. Information on the logic and operation used by the algorithms – the types of profiles created (if relevant), the variables used and whether they relate to personal information, the model parameters, detail about the training and validation data, detection of inaccuracies or errors, and audits or impact assessments.

3. Information on the consequences of using the algorithm – consequences of decisions for workers, male and female equality and the potential for biased outcomes.

4. Including any other relevant information – informing workers about the use of algorithms for automated decisions.

What’s next?

With regulatory efforts ramping up around the world, designers, developers, providers, and deployers of AI systems will soon see themselves faced with stringent obligations. With different countries taking slightly different approaches and targeting different applications of AI, businesses will need to act early to prepare to be compliant.

At Holistic AI, we have pioneered the fields of AI ethics and AI risk management and have carried out over 1000 risk mitigations. Using our interdisciplinary approach that combined expertise from computer science, law, policy, philosophy and ethics, and social science, we take a holistic approach to AI governance, risk, and compliance, ensuring that we understand both the technology and the context it is used in. To find out more about how Holistic AI can help you get compliant with upcoming AI legislation, get in touch at we@holisticai.com.