I'll tell you that we will talk about the boundaries of the future—practically without boundaries—artificial intelligence and its consequences for the world of work. I call Vincenzo Caridi, Head of Department at the Ministry of Labor; Alberto Gusmeroli, Member of Parliament, please; and Pietro Labriola, CEO of TIM.
So, we need to raise our voices, otherwise people in the back won't hear us. And since we're talking about artificial intelligence—which is something that isn't just about passion, because those who study it... We're going to turn off all your phones. The CEO of TIM is here. I don't know, I mean the risk is that, so let's not get him agitated. So, since artificial intelligence is no longer just study—artificial intelligence is now application, we've taken a step forward—I would start right with you, Mr. Labriola, because I'd like to understand. TIM is a huge company that will soon turn everything off, I want to tell you.
Here's the threat of turning off... Iliad's.
That one, absolutely. That deserves applause. Mr. Labriola, I have to wake them up like this; it's after lunch. In such a large company, is artificial intelligence an opportunity or not? And above all, what kind of application is there?
Certainly, artificial intelligence is a great opportunity to regain effectiveness and efficiency in both large and small-to-medium enterprises. The main problem in adopting artificial intelligence is still the difficulty within corporate culture. When someone starts talking about corporate culture in these contexts, it seems like we're discussing topics taken from university textbooks. However, to this day, despite everything that has been said, written, and heard, inside companies artificial intelligence is experienced by colleagues as a threat. In the sense that, even while immediately realizing that using AI can make an individual's job easier, that person naturally and humanly goes on the defensive regarding their job.
Let's try to be pragmatic and explain where artificial intelligence can act. Let's start with the most basic things: call centers. Today, whenever any of you have a problem, you speak with a—ugly to say—human operator. Well, that type of interaction, with the evolution of technology, becomes much simpler and easier with artificial intelligence. Some of you might still remember the early interactions when you tried talking to automated call centers, and there were those voices reminiscent of the Rome train station announcing the 12 o'clock train. I can guarantee you that using these services today makes it hard to distinguish a human being from artificial intelligence. We have a customer base that is somewhat older, and I can guarantee you that we have various recordings of customers who, while talking to artificial intelligence, say thank you and are convinced they are dealing with a human being.
Let me give a much simpler example. Who among you hasn't read or heard about telephone scams that took place? I recall the Ferrari case, where with artificial intelligence they cloned the voice of Ferrari's CEO calling the CFO asking him to sign things. So, from a customer experience standpoint, it's the same. What is the advantage? If I need to answer a million calls, in five minutes I can set up one million responses with artificial intelligence. If I have to answer a million calls with one million people, it's more complex. Also, from a training standpoint, if I want the answer to a customer's question to be standardized in a specific way, I have to train a million people, whereas with artificial intelligence, I can train just one model.
This is one area, along with all back-office procedures. Normally, you receive a customer complaint via email, or someone still sends it by fax; with artificial intelligence, you are able to process and work on that. This is the simplest part. But I think some of you have also heard about so-called AI agents. What are agents? We are now facing such an evolution where you have a tool at your disposal where you say, 'I need you to extract this data from system X, compare it with data in another system, check what works and what doesn't, and then upload it to another system.' Until now, doing this type of activity required either deploying a huge number of people, or what most companies did: developing new systems, which meant turning to software development companies to build a new system. Today that is no longer necessary because you can rapidly develop an agent.
What is the other side of the coin to all this, Mr. Labriola?
The other side of the coin, which we can also share here in our reasoning, is that certainly some jobs will be replaced. We cannot... I was giving an example earlier: it's like when you had a candle factory before public lighting and electricity arrived. That factory produced candles or lamps to illuminate streets and homes. If anyone thinks, 'Let's block the arrival of AI to protect jobs,' we are dead. What we need to understand is how to make things more flexible. And let me give some pointers. As President of Asstel, I negotiated the renewal of the collective labor agreement with the sector's trade unions. In this renewal, what condition did I set to sign it after four years of disputes? That I guarantee employment, but I do not guarantee the specific job role.
Let me explain better. If a job disappears, the union cannot force me to keep that job with that specific duty. If I used to produce candles and today I produce lightbulbs, you cannot ask me to keep having that person manufacture candles, because the company will fail and the job will be lost anyway. It sounds trivial, but these are constraints that, in the current context and in ongoing negotiations with social partners and unions, often become a problem.
Indeed. Let's give a round of applause to Mr. Labriola. Mr. Caridi, is it that simple, or is the matter a bit more complex?
Yes, good evening everyone and thank you for the invitation. Can you hear me? Yes, I'll speak louder. First of all, I was born in 1964, the year in which 1,060,000 babies were born. In 2025, according to the latest ISTAT data, only 355,000 babies were born. Why do I share this data? Because, as Mr. Labriola was saying, artificial intelligence can lend us a hand—a huge hand. It's not about losing jobs, absolutely not. Also because I'll tell you another figure: in more than 50% of positions, companies cannot find the skills necessary for their operations, because naturally artificial intelligence has changed the approach, it has changed the dimensions of the issue.
How do we tackle this approach? Obviously, if one million babies were born in '64 and in 2025 only 355,000 are born, we clearly have a problem acquiring skills. Earlier you mentioned this is very interesting for young people; artificial intelligence will be very interesting for mature individuals like me. Mature individuals, because it's not just about training young people, but even reskilling the competencies of those who already have a job, because today's job will no longer be what it is tomorrow.
And so, without being oracles, what is being done from a government perspective? Which, let me remind you, the Italian legislature was the first to implement the European AI Act with Law 132 of 2025, establishing an Observatory on Artificial Intelligence at the Ministry of Labor to examine the impact AI has on the world of work. I invite all of you to visit the Ministry of Labor website and consult an information system we called SIO, which is experimental and uses multi-agent generative artificial intelligence—what Mr. Labriola was talking about. That is an observatory that doesn't just provide dashboards or analyses and data—many observatories do that and do it well—but an observatory where each of us can understand, depending on our role in life (a young person, a new worker, a student, an older worker), how AI impacts our sphere of action.
This is important because it means giving people awareness of what will happen—not only the jobs that change, but also the skills needed to manage those jobs. Today, I repeat, the issue is not just a shortage of workers, but building skills. Multi-agent generative AI and the Ministry's systems... I really want to invite you to visit the website, use them, and provide feedback, because they are built on the feedback of people who tested this system. With these tools, we can finally provide guidance to everyone. It's not a generational issue. Today in companies, to conclude, we say there are five generations coexisting, but in reality, there are sectors that don't have this generational mix because there are more older workers than young ones, and who will transfer the skills of older workers? In this sense, artificial intelligence can help us enormously.
Thank you. This time a glance was enough. We are warming up, but unfortunately we only have seven minutes left, though we can manage another round. MP Gusmeroli, we have already seen examples of how artificial intelligence is applied in work. Mr. Labriola gave just a few examples, and we saw how the Ministry of Labor is operating. How do you regulate a sector that is so complex?
Beyond the European regulation, in just a few minutes I will share some brief highlights. One thing is artificial intelligence and how much it affects citizens versus small and medium-sized enterprises and large corporations. In the corporate world, about 70% of large companies have been using artificial intelligence systems for some time now. SMEs are still lagging behind. When we conducted the fact-finding inquiry on AI to help draft the law by the Ministry of Enterprises and Made in Italy, we saw it was at 16%; now it is around 20% of SMEs using AI systems—still few. But what we observed is that those using AI systems see a remarkable improvement in productivity and competitiveness.
So we cannot think of acting like those who want to stop water. You cannot stop water; you have to govern it. Artificial intelligence must be governed; it must be viewed with all its capacity for innovation and improvement in quality of life. But at the same time, precisely because it must be governed, we have to remember one figure: among citizens, 54% of Italians between 16 and 74 have no computer skills whatsoever. Imagine what the impact of artificial intelligence could be on citizens, more than half of whom currently have no ability to use IT tools. This is where legislation and politics must step in to protect those citizens so there is no social exclusion.
In the SME world, on the other hand, we must incentivize as much as possible because we face productivity issues caused by energy. We are scarce in energy sources until we increase renewables, until nuclear power arrives, and other sources—we won't easily increase hydro or thermal energy, which has massive growth paths. Nonetheless, energy is a huge problem for SMEs. We can improve the productivity and competitiveness of SMEs precisely by incentivizing these systems. However, we must pair incentives with commitments to maintain jobs. On this, I completely agree with the CEO of TIM: commitments to job maintenance must safeguard wages, certainly. But regarding specific job roles, we need to reflect, because if certain tasks are no longer performed and others must be done to preserve jobs, we must move in that direction in a balanced way. Parliament and policy can be decisive because SMEs need incentive support, and we have many initiatives underway and more we can still do.
