AI governance is failing because the people implementing it speak different languages. Lawyers don't understand the technical foundations of the systems they're regulating, and engineers don't grasp the legal frameworks constraining their work. A groundbreaking summer school in Italy just exposed this gap and offered a potential solution: forcing interdisciplinary teams to work together from the start.

What Happened at Europe's First AI Governance Summer School?

From June 10 to 12, 2026, Fondazione Bruno Kessler in Trento, Italy hosted the inaugural "AI + DATA: Practice & Law" summer school, a three-day intensive program designed to build practical skills in AI governance across legal, ethical, and technical domains. The event attracted 32 carefully selected participants from 81 applications, creating a deliberately diverse cohort that included lawyers, computer scientists, policy experts, and business leaders from universities, private companies, public administrations, and international organizations.

The program wasn't theoretical. Participants spent the first day learning the technical and legal context of AI systems, then moved into hands-on workshops exploring real governance challenges. The curriculum covered the EU's General Data Protection Regulation (GDPR), the EU AI Act, cybersecurity, intellectual property, risk assessment, and regulatory sandboxes. A particular focus went to the European Commission's "Digital Omnibus," a set of legislative proposals aimed at simplifying Europe's complex digital rulebook.

What made this summer school different from typical training programs was its explicit commitment to breaking down professional silos. The organizers recognized that AI governance requires people from different fields to collaborate effectively, and that collaboration is nearly impossible when they don't share a common vocabulary or understanding of each other's constraints.

Why Can't Experts in Different Fields Understand Each Other?

The core problem is straightforward but rarely addressed directly in policy circles: professionals working on AI governance operate in isolation. Lawyers need to understand the technical foundations of the systems they're writing rules for, but most legal training doesn't include computer science. Engineers building AI systems need to know the legal obligations they're creating, but most computer science programs don't teach data protection law or regulatory frameworks.

"One of the main challenges in AI governance is that people need to learn to speak the same language. In interdisciplinary work, a lot can get lost in translation: lawyers need to understand the foundations of the technologies they foster compliance with, just as computer scientists need to become familiar with the basic legal framework in which the technologies they develop operate. Nobody operates in a vacuum," explained Giulia Olivato, scientific coordinator of the summer school.

Giulia Olivato, Scientific Coordinator, AI + DATA Summer School

This communication gap has real consequences. When lawyers don't understand how machine learning models work, they write rules that are either too vague to enforce or technically impossible to implement. When engineers don't understand data protection law, they build systems that violate regulations without realizing it. The result is a patchwork of compliance efforts that don't actually address the underlying risks.

How Can Organizations Build Better AI Governance Teams?

The summer school's success offers practical lessons for companies and governments trying to improve their AI governance. The key is creating spaces where professionals from different disciplines can work together on real problems, not just attend separate lectures.

• Hire Interdisciplinary Teams: Don't separate your legal, technical, and ethics teams. Bring them together on AI governance projects from the beginning so they develop shared understanding of both constraints and possibilities.
• Invest in Cross-Training: Lawyers should learn enough about machine learning to understand what they're regulating. Engineers should understand data protection law and regulatory obligations. This doesn't require becoming an expert in the other field, but it requires moving beyond surface-level knowledge.
• Create Shared Vocabulary: Establish common definitions and terminology that work across disciplines. When a lawyer says "risk assessment" and an engineer says "risk assessment," they may mean completely different things. Clarifying these terms prevents costly misunderstandings later.
• Use Real-World Case Studies: Train teams using actual governance challenges they'll face, not abstract scenarios. The summer school's focus on practical implementation rather than theory proved far more valuable than traditional lectures.
• Support Ongoing Dialogue: Governance isn't a one-time training event. Build mechanisms for continuous communication between legal, technical, and ethics teams as regulations evolve and new AI systems emerge.

Who Attended and What Did They Bring to the Table?

The summer school's international and sectoral diversity strengthened the learning experience. While 74 percent of applicants came from European countries, 26 percent came from outside the EU, bringing perspectives on how different legal systems approach AI governance. Participants represented agriculture, banking, information technology, media, universities, nonprofits, and research institutions.

Particularly valuable were contributions from private sector professionals who are actively implementing AI governance in their organizations. Their real-world insights confirmed that the summer school's practice-oriented approach addressed genuine needs. These professionals could explain what governance frameworks look like when you're actually trying to deploy AI systems at scale, not just writing policy papers.

The Trento Bar Association's decision to award six professional training credits to lawyers who completed the program signals that legal professionals recognize the value of this interdisciplinary approach. For lawyers working on data protection, AI, and emerging technologies, understanding the technical and organizational context of AI systems is no longer optional.

What Does This Mean for AI Regulation Going Forward?

The strong response to the summer school, with 81 applications for just 32 spots, indicates growing demand for this type of training. As AI systems become increasingly embedded in public and private decision-making, professionals need practical tools and shared competencies to govern them responsibly.

The summer school's success also suggests that current approaches to AI governance may be missing something fundamental. Regulations like the EU AI Act are important, but they only work if the people implementing them can actually communicate with each other. A brilliant legal framework means nothing if the engineers building the systems don't understand it, and technical safeguards fail if lawyers don't know they exist.

The initiative was supported by Confindustria Trento, the Trento Bar Association, and the BioLaw Project at the University of Trento, reflecting recognition across research, professional, and business communities that building bridges between these groups is essential. This institutional support suggests that the summer school model could be replicated in other regions and countries facing similar governance challenges.

As AI systems continue to evolve and regulations become more complex, the ability of professionals from different fields to work together effectively will determine whether governance actually protects people or just creates bureaucratic obstacles. The summer school's core insight is simple but powerful: shared language and mutual understanding aren't nice-to-have extras in AI governance. They're foundational.