Artificial Intelligence and Corporate Governance, Stefano Nigro (Mindlogix): "The real challenge is control"

The global debate on the future of new technologies often shifts attention to distant, apocalyptic scenarios. But for Stefano Nigro, Chief Executive Officer of Mindlogix s.r.l., the reality is much more concrete and closer to companies’ everyday work: "Tegmark says that artificial intelligence is already out of control. In the companies we work with, I see a loss of control that is less spectacular, but much closer to home. Control of AI is not lost in a Silicon Valley laboratory: it is lost when a company connects an agent to its systems without deciding what it can do and who is accountable for it. And often without asking what data it will be able to read, where that data will end up and who else will be able to see it. These are questions that should come before the project, but instead come afterwards".

This initial area of concern is compounded by an even more hidden and widespread threat within organisations: "Then there is so-called shadow AI: employees using artificial intelligence tools on their own initiative, uploading company documents and data without the company knowing. In that case the issue is not even raised, because nobody knows that AI is already there".

Faced with an often uncontrolled technological race, Nigro also recalls the words of Pope Leo XIV, who noted that those who experience AI development as a race to be won risk looking only at the finish line, prioritising an immensely powerful and unpredictable system over one that is reliable and human-centred. According to the CEO of Mindlogix, this is a fundamental criterion that also guides companies’ strategic choices: while acknowledging that creating new governance institutions will take time, Nigro stresses that Europe already has a concrete tool. The AI Act establishes that the most sensitive systems must remain under effective human oversight, entrusted to competent people able to understand the machine’s decisions, intervene and, if necessary, stop it. This measure should not be read as mere bureaucracy, but as the dividing line between a system that is simply powerful and one that is truly reliable, as well as the only concrete way to maintain control.

In conclusion, the manager calls for companies not to wait for formal constraints or top-down decisions, but to embed a real culture of responsibility in everyday corporate management: "The point is to apply this principle every time an agent is put into production, even where the law does not require it. AI governance cannot remain in Brussels or in international forums: it must enter every organisation, starting with knowing which tools are being used, by whom and with what data".