The European Union’s Artificial Intelligence Act marks a watershed moment in technological governance, representing what Veale & Zuiderveen Borgesius (2021) term a “constitutional moment” for algorithmic regulation. As the European Parliament’s Rapporteur for the AI Act, Sandro Gozi’s insights illuminate the complex interplay between technological determinism and regulatory pragmatism that has shaped this  landmark legislation.

The Politics of Multi-stakeholder Governance and Technical Democracy

The Act’s development process reveals the inherent tensions in contemporary technology governance. “These are societal issues. Everybody felt involved,” Gozi observes, highlighting what Schaake & Barker (2020) identify as the “polycentric nature of AI governance.” The convergence of corporate interests, state sovereignty concerns, and civil society demands created a regulatory crucible that challenged traditional hierarchical approaches to lawmaking. This dynamic exemplifies what Bradford (2020) terms the “Brussels effect” in action, where EU regulatory power shapes global standards through market mechanisms rather than direct authority.

The varying levels of AI literacy among policymakers emerged as a critical epistemological challenge within this complex stakeholder landscape. The advent of ChatGPT during negotiations served as what Kaminski (2021) describes as a “technological disruption of regulatory assumptions.” This phenomenon underscores Hildebrandt’s (2020) argument that effective technological regulation requires a new form of literacy that bridges technical capability and democratic accountability. The challenge of maintaining adequate technical understanding while crafting broadly applicable legislation reveals what Pasquale (2020) terms the “expert-democracy tension” in technological governance.

Public Trust and Democratic Resilience in the Age of Algorithms

Gozi’s emphasis on public apprehension reveals a deeper philosophical tension in AI governance. The regulatory framework attempts to address what Pasquale (2020) terms the “black box society” phenomenon, where technological opacity breeds social anxiety.This dynamic illustrates Brkan’s (2021) observation that public trust in AI systems is  fundamentally linked to the legibility of their governance structures.

The Act’s provisions on democratic integrity reflect what Kaminski & Malgieri (2021) identify as the “constitutionalization of algorithmic accountability.” By mandating transparency and human oversight, the legislation acknowledges what Yeung (2020) describes as the “socio-technical nature of democratic processes.” Gozi’s concerns about foreign interference through AI systems exemplify what Brkan (2021) terms the”algorithmic manipulation of democratic discourse.” The legislation’s approach to these challenges demonstrates what Hildebrandt (2020) calls “democracy-preserving  innovation,” where technological advancement is deliberately shaped to reinforce rather than undermine democratic values.

The Innovation-Regulation Nexus and Data Governance

The Act’s risk-based approach represents what Ebers (2021) calls “graduated regulation,” a nuanced attempt to reconcile innovation with public protection. This framework exemplifies what Malgieri & Comandé (2020) identify as the “regulatory innovation paradox,” where the speed of technological development challenges traditional regulatory timeframes. The centrality of data access in the Act highlights what they term the “data-knowledge nexus” in AI development, grappling with what Hildebrandt (2020) describes as the “data commons dilemma.”

The legislation’s approach to data governance reflects a sophisticated understanding of what Bradford (2020) terms the “data sovereignty challenge.” By establishing clear frameworks for data access and usage, the Act attempts to balance what Pasquale (2020) identifies as the competing imperatives of innovation and privacy. This balance is particularly evident in provisions regarding AI training data, where the Act seeks to establish what Yeung (2020) describes as “data governance ecosystems” that support both technological advancement and public interest.

Global Influence and Sociotechnical Equity

The EU’s aspiration to influence global AI governance reflects what Bradford (2020) terms “regulatory export through market mechanisms.” This approach exemplifies what Schaake & Barker (2020) identify as “normative power Europe” in the digital age, where regulatory standards become de facto global requirements through market access conditions. The Act’s attention to territorial, educational, and generational gaps acknowledges what Pasquale (2020) terms the “stratification of technological access.”These disparities represent what Yeung (2020) identifies as “regulatory blind spots” in technological governance, where formal equality can mask substantive inequities.

The global impact of the Act is further complicated by what Malgieri & Comandé (2020) describe as the “regulatory competition paradigm,” where different jurisdictions compete to set global standards for AI governance. This dynamic is particularly evident in the EU’s relationship with other major technology powers, where the Act serves as what Bradford (2020) terms a “regulatory first mover” in establishing comprehensive AI governance frameworks.

Adaptive Governance and Future Trajectories

Gozi’s acknowledgment of the need for future revisions reflects what Yeung (2020) terms “anticipatory regulation.” This approach embodies what Hildebrandt (2020) describes as “legal protection by design,” where regulatory frameworks are deliberately constructed to evolve with technological change. The Act’s flexible structure acknowledges what Brkan (2021) identifies as the “temporal challenge” in technology regulation, where governance frameworks must balance current certainty with future adaptability.

This adaptive approach is particularly evident in the Act’s treatment of emerging AI applications, where it establishes what Kaminski (2021) terms “regulatory sandboxes” for testing new governance approaches. The legislation’s forward-looking elements demonstrate what Pasquale (2020) describes as “regulatory foresight,” where governance frameworks anticipate rather than merely respond to technological change.

Conclusion

The EU AI Act represents more than mere regulation; it embodies what Bradford (2020) terms a “regulatory philosophy” that seeks to reconcile technological innovation with democratic values. As Gozi’s insights reveal, the challenge lies not just in technical rule-making but in what Pasquale (2020) identifies as the “socio-technical contract” between innovation and public good. The Act’s success will ultimately depend on its ability to fulfill what Yeung (2020) describes as the “regulatory promise” of ensuring that AI development serves rather than subverts democratic society.

References

• Bradford, A. (2020). The Brussels Effect: How the European Union Rules the World. Oxford University Press.
• Brkan, M. (2021). AI-supported Decision-making under the General Data Protection Regulation. International Data Privacy Law, 11(1), 37-56.
• Ebers, M. (2021). Regulating AI and Robotics: Ethical and Legal Challenges. In Artificial Intelligence and Robotics in the European Union: Opportunities and Challenges.
• Hildebrandt, M. (2020). Law for Computer Scientists and Other Folk. Oxford University Press.
• Kaminski, M. E., & Malgieri, G. (2021). Algorithmic Impact Assessments under the GDPR:  Producing Multi-layered Explanations. International Data Privacy Law, 11(2), 125-159.
• Malgieri, G., & Comandé, G. (2020). Why a Right to Legibility of Automated Decision-Making Exists in the General Data Protection Regulation. International Data Privacy Law, 10(1), 24-34.
• Pasquale, F. (2020). New Laws of Robotics: Defending Human Expertise in the Age of AI. Harvard University Press.
• Schaake, M., & Barker, T. (2020). Democratic Source Code for a New U.S.-EU Tech Alliance. Brookings Institution.
• Veale, M., & Zuiderveen Borgesius, F. (2021). Demystifying the Draft EU Artificial Intelligence Act. Computer Law Review International, 22(4), 97-112.
• Yeung, K. (2020). Regulation by Design: Towards a Regulatory Future for AI. European Journal of Law and Technology, 11(2), 1-23.