TLDR: This paper explores the complex relationship between AI regulation and innovation, arguing against a false dichotomy. It proposes a new approach to regulation that prioritizes fundamental rights, fosters responsible innovation, and establishes global accountability standards. The author suggests leveraging ethical and legal ecosystems, promoting AI literacy, and empowering users and educators to co-create regulatory frameworks from the ground up, rather than relying solely on top-down enforcement. The paper also highlights the geopolitical tensions and the need for a critical understanding of AI’s impact on society and education.
In an era where artificial intelligence (AI) is rapidly transforming every aspect of society, the debate between fostering innovation and ensuring robust regulation has often been framed as a difficult choice. However, a recent essay by Pompeu Casanovas, titled “On the false election between regulation and innovation. Ideas for regulation through the responsible use of artificial intelligence in research and education,” challenges this very notion, arguing that a balanced and responsible approach is not only possible but essential. This insightful paper, presented at the AI Hub-CSIC / EduCaixa Summer School, delves into critical questions surrounding AI governance, fundamental rights, and the future of international cooperation in AI development. You can read the full paper here.
Challenging the Dichotomy: Regulation and Innovation Hand-in-Hand
Casanovas begins by highlighting the “false choice” between digital regulation and innovation, a concept championed by international law expert Anu Bradford. Bradford’s “Brussels Effect” suggests that the European Union’s stringent data protection regulations, like GDPR, often become de facto global standards because tech giants cannot afford to lose the European market. The paper argues that Europe should continue to strengthen its enforcement power, prioritizing human and civic rights while still incentivizing technological advancement.
The essay points to recent legal battles over AI and intellectual property, such as the Bartz et al v. Anthropic and NY Times v. OpenAI lawsuits, as evidence of the urgent need for clear regulatory frameworks. It also references the European Court of Justice’s ruling against AI-based credit decisions by SCHUFA, reinforcing the principle that AI should be an assistive tool, not a sole decision-maker, especially when fundamental rights like privacy and non-discrimination are at stake.
Prioritizing Fundamental Rights in AI Development
To effectively prioritize fundamental rights without stifling innovation, Casanovas proposes a shift in how we conceive and construct regulatory models. This involves moving beyond traditional “Compliance by Design” (CbD) to “Compliance through Design” (CtD), integrating legal governance mechanisms directly into AI systems. Key preconditions for this include:
- Controlling the behavior and ideology of tech corporate leaders.
- Clearly defining the scope and implementation of normative systems.
- Building real-time ethical and legal ecosystems.
- Re-centering ethics in regulatory models, independent of specific jurisdictions.
- Fostering citizen proactivity in accepting and maintaining these ecosystems.
- Acknowledging and managing geopolitical conflicts, particularly those involving the US, Russia, and China.
The paper also critically examines the “luxury beliefs” of influential tech CEOs like Peter Thiel and Alexander Karp, suggesting that their social and political theories, often rooted in aggressive expansionism, pose a significant challenge to responsible AI development. It warns against an “oligarchy of knowledge” that could prioritize the law of the strongest in innovation.
Fostering Responsible Innovation Amidst Risks
Addressing the inherent risks of AI—such as bias, mass surveillance, manipulation, copyright infringement, disinformation, and mental health impacts—the paper advocates for a multi-faceted approach. It suggests leveraging “civic technologies” (like Ushahidi, used for crisis mapping) and developing ethical and legal ecosystems, particularly within sectors like Industry 5.0 (smart manufacturing), where standards can be built from the ground up.
Casanovas introduces the concept of “legal over-compliance,” where entities formally comply with regulations, sometimes excessively, to protect investments while potentially diverting attention from ethical obligations. He emphasizes the importance of the “AI value chain” and AI literacy, as outlined in the EU’s AI Act (2024), to ensure that AI systems are developed and deployed responsibly, with their utility and acceptance by users determining their true value.
Global Standards for AI Responsibility
When considering international cooperation, especially given the US’s flexible approach to AI regulation, the paper argues against a “race to the bottom” in rights. Instead, it calls for mechanisms that promote self-regulation, co-regulation, and value creation. Examples include AI sandboxes for legal experimentation and the Markets in Crypto-Assets Regulation (MiCA), which introduces decentralized authority into financial markets.
The paper contrasts the regulatory strategies of major global players: the EU’s comprehensive AI Act, the UK’s principle-based “pro-innovation” approach, and the US’s sectoral “patchwork” model. It notes that even within the EU, there’s a call for simplification of regulations like GDPR to foster a more competitive digital market. The core argument is that responsible regulation cannot be imposed solely from the top down but must emerge from the proactive engagement of users, designers, and creators at the micro and meso levels.
Also Read:
- Unpacking “Accountability Capture” in AI Oversight: The Unseen Impacts of Record-Keeping
- Navigating AI: How Mission-Driven Organizations Adopt Technology with Purpose
Implications for Education and Research
A significant portion of the essay is dedicated to the implications for education. Casanovas stresses the need for educators and researchers to have a precise understanding of AI development conditions and the global geopolitical landscape. He highlights the vast and growing field of AI in Education (AIED), acknowledging both its potential benefits and limitations, and the critical ethical dimension.
The paper references the ODITE Report (2025) on personalized learning and the “Manifesto in Defence of Human-Centred Education in the Age of Artificial Intelligence” (2025), which proposes six levels of creative participation in human-AI interaction in education. This framework moves from passive consumption to expansive learning and co-creation, emphasizing that AI literacy is crucial for preserving human agency. The author suggests that educational institutions can lead by example, developing concrete, local, and regional frameworks for responsible AI use, driven by the real-world experiences of students, teachers, and researchers.
Ultimately, the paper concludes that responsible innovation and public interest are not mutually exclusive but can be harmonized through the active involvement of all stakeholders. It calls for a critical yet proactive approach, where the complexities of current regulatory frameworks are understood, and new, collaborative models are forged from the ground up to ensure AI serves humanity responsibly.
