Governing AI before AI governs us: MiRò at ExpoAid 2026

On June 26th, on the stage of the Sala Bastioni at the Palacongressi di Rimini, a discussion took place on an issue in communication that can no longer be postponed: whoever decides how we talk about artificial intelligence also decides how we govern it.

This is the theme at the heart of the address given by Rossella Rosciano, CEO of MiRò Comunicazione, at the round table organized by the Lions of Multidistrict 108 as part of ExpoAid 2026, moderated by Rossella Vitali, President of the Council of Governors of MD108. Alongside her were Andrea Cornelli, Flavio Ceravolo, and Massimo Banzi, bringing together academic, professional, and civic perspectives on an issue that concerns everyone, even before it concerns technology itself: power.

An uncomfortable starting point

Artificial intelligence isn’t waiting for institutions to be ready. It is already reshaping access, power, and responsibility, often without civil society having any say in the matter. This is the starting point of Rossella’s reflection, built on a precise theoretical framework: Edward Freeman’s stakeholder theory.

An entity is considered a stakeholder when it has a recognizable interest, the power to influence decisions, and legitimacy acknowledged by the system. Today, AI meets all three criteria: it has its own interests in optimizing its objectives, concrete power over economic and relational decisions, and growing legitimacy, implicitly sanctioned even by the European AI Act. The conclusion is clear: if AI is a stakeholder, it must be managed as an actor, no longer treated as a mere tool.

Three narratives that need rethinking

At the core of the address are three ways in which the public narrative around AI is already exercising a form of governance, without democratic mandate or accountability.

The first is the selective risk narrative: media amplify what is visible and dramatic, such as deepfakes and the automation of manual labor, while leaving systemic risks in the shadows, like algorithmic bias, which today already influences access to credit, healthcare, and justice.

The second is the appropriated opportunity narrative: the story of AI as a democratic and inclusive tool is produced almost exclusively by those who develop and sell it. There is little to no counterbalance built on independent evidence.

The third, perhaps the most significant, is the absent responsibility narrative: when AI makes a mistake, accountability systematically dissolves between the algorithm, the training data, the base model, and the end user. This accountability gap isn’t an accident, but the result of precise narrative choices.

The role civil society must play

The proposal emerging from the round table isn’t consultative in nature. It isn’t about “participating in the debate on AI” — a phrasing that, according to Rossella, is already a narrative trap, since it presupposes a table set by others. Instead, it’s about exercising oversight: monitoring dominant narratives, mapping cases of unattributed harm, and building AI literacy tools for citizens and local administrators. Initiatives like the Permanent Observatory promoted by the Lions MD108 go exactly in this direction.

Why this matters for communication professionals too

For those working in communications, the relevant question isn’t “how do I use AI in my work,” but what role the profession plays in making AI understandable, accessible, and governable even for those who don’t build it. It’s a responsibility MiRò chooses to carry forward beyond its own direct projects, by taking part in the public debate on issues that will define the boundaries of communication in the years to come.

The window to shape how AI will be governed is still open. But, as Rossella Rosciano reminded the audience at the close of her address, it is closing: every day without an authoritative, independent civic narrative is a day in which AI governance is exercised by default — by those who hold no mandate to do so.