Spain - Catalonia - Ethical AI Framework

Overview

The concept of the 'Catalonia Office of Ethical and Responsible AI' is primarily embodied by the Observatory for Ethics in Artificial Intelligence of Catalonia (OEIAC), an integral part of the broader Catalonia.AI Strategy launched by the Government of Catalonia in 2020. This initiative underscores Catalonia's commitment to fostering an artificial intelligence ecosystem that is not only innovative and economically beneficial but also deeply rooted in ethical principles and respect for fundamental rights. The OEIAC, often taking the form of a Chair at the University of Girona, serves as a crucial advisory and research body, dedicated to exploring the multifaceted ethical, social, and legal implications arising from the deployment of AI technologies in daily life across Catalonia. Its establishment reflects a proactive stance by the Catalan government to ensure that AI development and application align with societal values and legal frameworks, particularly in anticipation of and alignment with European Union regulations such as the AI Act.

The OEIAC's mission extends beyond mere academic study; it actively promotes the implementation of ethical and responsible AI practices through tangible tools and guidelines. A cornerstone of its efforts is the development and dissemination of the PIO (Principles, Indicators, and Observables) self-assessment model, a pioneering instrument designed to assist both public and private entities in evaluating their AI systems for compliance with ethical standards and existing regulations. This proactive approach by Catalonia aims to position the region as a leader in trustworthy AI, ensuring that technological advancements are compatible with human-centric values and contribute positively to the well-being of its citizens. The work of the OEIAC is intrinsically linked to the wider ambition of the Generalitat de Catalunya to navigate the complexities of AI development responsibly, fostering an environment where innovation thrives without compromising ethical integrity or societal trust.

Definitions

The Observatory for Ethics in Artificial Intelligence of Catalonia (OEIAC) is a pivotal entity within Catalonia's AI landscape. It functions as a specialized body, often structured as a Chair at the University of Girona, dedicated to analyzing the ethical, social, and legal ramifications of artificial intelligence. Its core mandate involves promoting the responsible and ethical application of AI systems, ensuring they adhere to prevailing legal standards and international best practices. The OEIAC is strategically placed within the 'Ethics and Society' axis of the Catalonia.AI Strategy, reflecting its foundational role in guiding the region's AI development towards human-centric and trustworthy outcomes.

The Catalonia.AI Strategy represents the overarching Artificial Intelligence Strategy of the Government of Catalonia. Launched in 2020, this comprehensive plan outlines a strategic vision for enhancing Catalonia's AI ecosystem. Its objectives encompass fostering economic growth, improving the quality of life for citizens, and crucially, ensuring the ethical deployment of AI technologies. The strategy is structured around six key lines of action, providing a holistic framework for AI development, research, talent cultivation, infrastructure, adoption, and ethical considerations. The OEIAC is explicitly identified as a fundamental pillar of this strategy, underscoring the central role of ethics in Catalonia's AI ambitions.

The PIO Self-assessment Model (Principles, Indicators, and Observables) is an innovative and publicly accessible tool developed by the OEIAC. This model serves as a practical checklist for organizations to evaluate their AI data and systems against established ethical principles and legal obligations. By providing a structured verification process, the PIO model helps identify appropriate or inappropriate actions in AI development and deployment, thereby promoting compliance with AI risk regulations. It is designed to raise awareness across the 'quadruple helix' (administration, academia, business, and citizens) regarding the ethical and responsible use of AI, aligning with international standards for trustworthy AI.

Additionally, the Direcció General d'Intel·ligència Artificial, Eficiència i Dades a l'Administració (Directorate General for Artificial Intelligence, Efficiency and Data in the Administration) is a newly established governmental body within the Generalitat de Catalunya. Approved in May 2025, this Directorate General is set to become operational before the summer of the same year. Its purpose is to lead and coordinate the digital innovation strategy across the administration, with a specific focus on promoting the ethical, responsible, and secure integration of AI and advanced data utilization to enhance the quality, efficiency, and personalization of public services. This formal governmental office complements the advisory role of the OEIAC by directly implementing AI policies within the public sector.

Governance and Institutional Framework

The governance structure for ethical and responsible AI in Catalonia is multifaceted, with the Observatory for Ethics in Artificial Intelligence of Catalonia (OEIAC) serving as a central, albeit advisory, institution. The OEIAC is formally integrated into the Artificial Intelligence Strategy of the Government of Catalonia, known as Catalonia.AI, which was launched in 2020. Within this strategy, the OEIAC is positioned under the 'Ethics and Society' axis, highlighting its foundational role in guiding the ethical development and deployment of AI across the region. The Observatory often takes the form of a Chair at the University of Girona, fostering a strong link between academic research and public policy in the field of AI ethics. This institutional arrangement allows the OEIAC to leverage academic expertise to study the complex ethical, social, and legal implications of AI, thereby informing and shaping Catalonia's approach to AI governance.

Complementing the OEIAC's advisory role, the Generalitat de Catalunya has further solidified its institutional framework for AI governance with the recent approval of the Direcció General d'Intel·ligència Artificial, Eficiència i Dades a l'Administració (Directorate General for Artificial Intelligence, Efficiency and Data in the Administration) in May 2025. This new Directorate General, slated to be operational before the summer of 2025, represents a more direct governmental body focused on implementing AI policies within the public administration. It is designed to coordinate, lead, and execute the digital innovation strategy, explicitly promoting the ethical, responsible, and secure use of AI. This development signifies a robust commitment by the Catalan government to embed ethical AI considerations directly into its administrative operations, ensuring that the principles advocated by the OEIAC are translated into practical governmental policy and action. The combined efforts of the OEIAC and this new Directorate General establish a comprehensive governance framework aimed at guiding AI development in Catalonia towards responsible and human-centric outcomes.

Key Focus Areas

The core focus of the Catalonia Office of Ethical and Responsible AI, primarily driven by the OEIAC, revolves around a set of fundamental ethical principles designed to ensure that AI development and deployment in Catalonia are trustworthy and beneficial to society. These principles, which form the basis of the OEIAC's assessment tools and recommendations, include: Transparency and Explainability, aiming to ensure that AI operations and decision-making processes are clear, understandable, and interpretable by users and stakeholders; Justice and Fairness, to actively prevent biases and guarantee equitable treatment and outcomes for all individuals interacting with AI systems; Safety and Non-maleficence, to mitigate potential harms to users and society that could arise from unethical or irresponsible AI development or use; and Responsibility and Accountability, establishing clear guidelines for accountability regarding actions and decisions made by or with AI systems, thereby ensuring responsible development and deployment. These principles are crucial for building public trust in AI technologies.

Further expanding on its ethical mandate, the OEIAC also emphasizes Privacy, prioritizing the robust protection of personal and sensitive data, upholding privacy rights, and maintaining confidentiality, particularly for end-users of AI systems. Autonomy is another key focus, dedicated to safeguarding and promoting human agency in contexts where AI introduces automation, ensuring that individuals retain meaningful control over AI systems. Lastly, Sustainability is highlighted to protect the environment and consider the ecological and social impacts of AI, advocating for sustainable practices and long-term viability in AI development. By focusing on these seven interconnected principles—Transparency, Justice, Safety, Responsibility, Privacy, Autonomy, and Sustainability—the OEIAC provides a comprehensive ethical framework that guides the ethical assessment of AI data and systems, aligning Catalonia's AI strategy with global ethical standards and human rights considerations.

Implementation Framework

The implementation framework for ethical and responsible AI in Catalonia, largely spearheaded by the OEIAC, centers on practical tools and guidelines designed to integrate ethical considerations into the entire lifecycle of AI systems. A cornerstone of this framework is the innovative PIO (Principles, Indicators, and Observables) self-assessment model. This model provides a structured, open-access mechanism for organizations—ranging from public administration to private enterprises—to systematically evaluate their AI data and systems. It functions as a comprehensive checklist, enabling users to verify compliance with both existing AI regulations and the ethical principles advocated by the OEIAC. The PIO model is instrumental in identifying potential ethical risks and non-compliant actions, thereby fostering a culture of accountability and responsible AI development.

Beyond self-assessment, the OEIAC's implementation framework also includes the provision of ethical recommendation clauses for the procurement of AI systems. These clauses serve as vital guidelines for public and private organizations, encouraging them to embed ethical considerations directly into their purchasing and contracting processes for AI technologies. By offering these recommendations, the OEIAC aims to influence the market towards the development and adoption of ethically sound AI solutions from the outset. This proactive approach ensures that ethical principles are not merely an afterthought but are integrated into the foundational stages of AI acquisition and deployment. The framework's emphasis on both internal self-regulation through the PIO model and external influence through procurement guidelines demonstrates a holistic strategy for embedding ethical and responsible AI practices across Catalonia's digital ecosystem.

Monitoring and Evaluation

The monitoring and evaluation mechanisms for ethical and responsible AI in Catalonia are primarily facilitated through the OEIAC's pioneering PIO (Principles, Indicators, and Observables) self-assessment model. This model serves as a robust evaluation tool that enables a comprehensive verification process for AI data and systems. Its design allows for the identification of actions that are either appropriate or inappropriate in relation to established ethical principles and current AI regulations. By providing a structured framework for assessment, the PIO model empowers organizations to continuously monitor their adherence to ethical standards, thereby promoting ongoing compliance and responsible innovation. This continuous evaluation process is crucial for adapting to the rapidly evolving landscape of AI technologies and their associated ethical challenges.

Furthermore, the PIO self-assessment model is not merely a compliance checklist but also a tool for raising awareness and fostering a deeper understanding of ethical AI across various stakeholders. It is designed to sensitize the 'quadruple helix'—comprising administration, academia, business, and citizens—to the importance of ethical and responsible uses of AI data and systems. Through this evaluative process, the OEIAC aims to cultivate a shared commitment to ethical AI development and deployment throughout Catalonia. The insights gained from these self-assessments can also inform policy development and guide future initiatives, ensuring that Catalonia's AI strategy remains responsive to emerging ethical concerns and aligned with the highest standards of accountability, safety, and respect for human rights.

Penalties, Liability, and Appeals

The Catalonia Office of Ethical and Responsible AI, primarily through the OEIAC, operates as a guiding and advisory body rather than a direct enforcement agency with powers to impose penalties or determine liability. Its role is to promote compliance with existing and forthcoming AI regulations, such as the EU AI Act, and to foster ethical AI development through education, guidelines, and self-assessment tools. Therefore, the OEIAC itself does not directly prescribe penalties, liability frameworks, or appeal processes. Instead, it aims to prevent issues by encouraging adherence to ethical principles and legal obligations, thereby reducing the likelihood of violations that would trigger penalties under broader national or European legislation. The emphasis is on proactive ethical integration rather than reactive punitive measures.

Any penalties, liability determinations, or appeal mechanisms related to AI systems in Catalonia would fall under the purview of existing national and European legal frameworks. For instance, the upcoming EU AI Act, once fully in force, will establish clear rules regarding high-risk AI systems, including provisions for enforcement, sanctions for non-compliance, and mechanisms for redress. The OEIAC's work, particularly its PIO self-assessment model and ethical recommendation clauses, is designed to help organizations navigate these complex regulatory landscapes and ensure they meet their obligations, thereby avoiding potential legal repercussions. In this context, the OEIAC acts as a facilitator for responsible AI governance, supporting entities in their efforts to develop and deploy AI in a manner that respects legal requirements and ethical standards, ultimately contributing to a more compliant and trustworthy AI ecosystem.

Relationship to Other Instruments

The Catalonia Office of Ethical and Responsible AI, primarily represented by the OEIAC, is deeply embedded within and intricately linked to several key instruments and strategies at both regional and European levels. Fundamentally, the OEIAC is a core component of the comprehensive Catalonia.AI Strategy, launched by the Generalitat de Catalunya in 2020. This overarching strategy provides the strategic context and mandate for the OEIAC's work, positioning ethical and responsible AI as a central pillar of Catalonia's digital transformation. The OEIAC's activities, including the development of the PIO self-assessment model and ethical guidelines, directly support the objectives of the Catalonia.AI Strategy to foster a trustworthy and human-centric AI ecosystem.

Furthermore, the work of the OEIAC and the broader Catalonia.AI Strategy are closely aligned with and responsive to the evolving regulatory landscape at the European Union level, particularly the EU AI Act. The Catalan government explicitly states its preparedness to apply the regulations derived from the EU AI Act, with the OEIAC's tools designed to facilitate compliance with these new European standards. The OEIAC's ethical principles, such as transparency, fairness, and accountability, resonate strongly with the foundational requirements of the EU AI Act, which aims to harmonize AI rules across the EU. In a more direct governmental development, the Generalitat de Catalunya also approved the creation of the Direcció General d'Intel·ligència Artificial, Eficiència i Dades a l'Administració in May 2025. This new Directorate General will play a crucial role in implementing ethical and secure AI within the public administration, building upon the foundational work and principles championed by the OEIAC and aligning with the broader regional and European commitments to responsible AI.

International Alignment

The Catalonia Office of Ethical and Responsible AI, through the work of the OEIAC, demonstrates a strong commitment to international alignment in the field of AI ethics and governance. The OEIAC's foundational ethical principles—Transparency, Justice, Safety, Responsibility, Privacy, Autonomy, and Sustainability—are explicitly designed to align with and integrate high-level international standards and ethical principles for the design, implementation, and use of AI systems. This global perspective ensures that Catalonia's approach to AI is not insular but contributes to and benefits from the broader international discourse on trustworthy AI. The Observatory's self-assessment tools and recommendations are crafted to be compatible with a growing international consensus on what constitutes ethical and responsible AI development.

A significant aspect of this international alignment is Catalonia's proactive engagement with and preparation for the European Union's AI Act. The Catalan government has emphasized its readiness to implement the forthcoming European regulations, recognizing the EU AI Act as a pioneering global standard for AI regulation. The OEIAC's efforts to promote compliance with AI risk regulations and to provide tools like the PIO self-assessment model are directly aimed at helping entities in Catalonia adapt to these new legal requirements. This strategic alignment with the EU AI Act positions Catalonia as a key player in the European vision for responsible AI, fostering cross-border cooperation and contributing to the harmonization of AI standards across the continent. By embracing international principles and regulations, Catalonia seeks to ensure that its AI ecosystem remains competitive, trustworthy, and globally responsible.

Implementation Timeline

Compliance Checklist

Sources and References