Episode 06 · september 2026

AI and digital trust: the technology is running ahead of the people using it

Half of Italians already use generative AI, only a third feel able to use it well. We examine what happens when a technology runs ahead of the awareness needed to govern it — from personalisation that slides into profiling to data security, digital literacy and rules that struggle to keep pace.

Around the table

Guests & host

Benedetta De Michelis

Manager Business Development, B2You Altroconsumo

Benedetta De Michelis is Business Development Manager at B2You Altroconsumo, the B2B service line that supports corporate partners in developing and refining products and services able to meet the needs of an increasingly informed consumer base — one that plays a growing role in shaping market dynamics. She joined Altroconsumo in 2001 and has worked throughout on building partnerships and services designed to respond to what consumers need and to make their choices easier.

Gabriele Ratti

CEO Moov · Host

CEO of Moov and host of Moov(er)Time, he guides each conversation toward the practical questions behind the headlines, keeping the focus on what changes for the people who actually move.

About this episode

What we talked about

AI didn’t wait for us to be ready. In this episode we start from an uncomfortable pair of numbers: 48% of Italians already use generative AI at least occasionally, while only 33% feel competent enough to use it well — a figure that collapses to 19% among over-60s no longer in work. We unpack what happens when adoption outruns awareness, why 66% of consumers fear the manipulation of public opinion while 58% worry about their data, and where personalisation stops being a service and starts becoming profiling.

We then turn to the companies building on top of all this. From hybrid models where AI assists people without replacing them, to collecting only the data a pay-per-use service genuinely needs, the conversation asks what changes when ethics stops being a compliance exercise and becomes the founding principle of a business decision — and how regulation and digital literacy can move together instead of leaving people behind.

Key themes

Highlights from the conversation

🔹 Why adoption is outrunning awareness — and what that gap costs when AI is already part of everyday choices

🔹 The skills divide that no infrastructure can close, and why it risks turning AI into a barrier rather than an opportunity

🔹 Where personalisation ends and profiling begins: what consumers should know about the data behind a tailored service

🔹 The hybrid model in practice: how AI can assist a team without quietly taking the decision out of human hands

🔹 Why ethics belongs inside a business decision — and how regulation and digital literacy only work when they move together

Full transcript

Read the full transcript

‍↓

How widespread is generative AI among Italians today, and how aware are they of how they use it?

‍
AI has made its way into everyday life and is no longer a tool used only by a narrow group of experts. According to an Altroconsumo survey, 48% of Italians use generative AI tools at least occasionally, while only 33% believe they have the skills to use them properly. The figures point to very rapid adoption, paired with a level of competence and awareness that lags well behind.

‍In which areas of daily life do Italians notice AI the most?

‍
People perceive AI across a wide range of areas: news and media, shopping, culture, entertainment, education, personal and professional communication, health and wellbeing. That perception is rising everywhere, and most visibly in communication and in health and wellbeing.

‍Is there a skills gap when it comes to using AI?

‍
Yes. The 33% of Italians who feel adequately skilled conceals a significant gap between groups. Among younger people the figure reaches 56%, while among over-60s who are no longer working it falls to 19%. This is a digital divide driven by skills rather than infrastructure. The risk is that AI tools become a lever of discrimination in access to opportunities and services, unless they are accompanied by proper training and awareness-building.

‍How does Italy compare with other European countries on skills?

‍
The same survey was run in nine other European countries, where an average of 44% of people believe they have sufficient skills, against 33% in Italy. That gap suggests Italy is behind on the training and awareness needed to use these tools and draw full benefit from them.

‍Is user satisfaction with AI output increasing?

‍
55% of respondents say they are satisfied with the results they get, but the figure is down on 2024. That may reflect a sharper awareness of the limits of these tools. The initial enthusiasm now comes with the understanding that AI does not necessarily deliver a true or definitive answer, and should be used critically. AI should not be treated as the only source of information or creativity: human judgement remains essential.

‍What approach can a company take to using AI?

‍
One option is a hybrid approach, where automation, machine learning and AI assist people rather than replacing them outright. Automation can improve certain activities, but it has to stay under the control of a team. In customer care, for example, an AI agent can handle informational questions while always leaving the door open for a person to step in. The same principle applies across the service, from sales through to after-sales and follow-up.

‍Is AI also changing the way consumers communicate with companies?

‍
Yes. Some customers are using AI tools to structure their requests to companies. Communications become more articulate and precise, with requests framed in a more organised way, which can make the exchange more effective because it gets to the point faster. From the company’s side, analysing those requests also helps pinpoint more clearly where customers most need information or clarification.

‍What worries Italians most about AI?

‍
Among the figures cited during the webinar, 66% of consumers are concerned about the risk of public opinion being manipulated, while 58% are concerned about privacy and data security. Consumers recognise both the opportunities and the risks of AI’s spread. On one hand they value more personalised content, experiences and recommendations made possible by data analysis; on the other, they fear the same data could be used for profiling or for purposes that are not sufficiently transparent.

‍Where is the line between personalisation and non-transparent use of data?

‍
It comes down to understanding what happens to the data you hand over when you use a product or service. AI can draw on user information to provide personalised answers, services or recommendations. But consumers need to be able to understand what data is collected, what it is used for and what choices they have about that use. Awareness is what allows a person to decide whether to favour the benefits of personalisation or take a more cautious approach.

‍Can AI be used for fraud?

‍
Yes. The information available about consumers can also be used to build increasingly personalised scams. One example is the use of tools capable of imitating the voice of someone the victim knows, making them believe they are speaking to a family member in trouble. AI can therefore be used not only to deliver new services, but also by people who intend to exploit it dishonestly.

‍Can AI influence or manipulate public opinion?

‍
The risk exists. AI systems can produce answers that steer users towards particular interpretations, and can contribute to a flattening of perspectives. The issue also extends to the growing volume of altered or artificially generated images, audio and video, which makes it harder to tell authentic information from constructed content. It is important to understand that AI output is not necessarily neutral: behind the result sit systems and models that can shape the kind of answer produced.

‍How is AI’s impact on work perceived today?

‍
In the survey cited, 68% of respondents perceive no impact of AI on their own job. That should be read as a snapshot of perception, not necessarily as a measure of AI’s actual impact on the labour market. In some areas, though, AI is already changing how people work and can raise productivity — particularly when it is used by people with the skills to understand both the tool and the results it produces.

‍Is AI replacing certain professions?

‍
At present there is no sign of a complete, widespread replacement of professions or processes. Some activities, though, are changing profoundly. In software development, for instance, new models and tools are shifting part of the work away from writing code and towards analysing what the software is meant to achieve. AI can therefore raise productivity in some activities while, at the same time, requiring the skills behind them to evolve.

‍Why should ethics be part of corporate decision-making rather than an external constraint on the business?

‍
Ethics in the use of data did not begin with AI. It has long concerned privacy, data use and respect for the purposes data was collected for. AI amplifies the scale of the problem, because it makes it possible to use and process vast quantities of information and can expose consumers to uses of their data that go beyond what they immediately perceive. That is why a company has to consider not only what the law permits, but also the principles of fairness it chooses to adopt towards consumers.

‍How does that principle apply to data collection in a pay-per-use service?

‍
The principle is to collect only the information needed to deliver the service and fulfil the stated purpose. In a pay-per-use service based on distance travelled, for example, the mileage figure is necessary to determine the price. The customer can see that figure in the app and, where provided for, interact with it to correct it. The same principle applies to behavioural data used to calculate particular indicators: the customer has to know what information is collected and why.

‍What does putting the consumer at the centre of data use actually mean?

‍
It means building a transparent relationship between the data collected, the purpose of the service and the benefit returned to the customer. Consumers must know what information is being used and why. They must also be able to exercise the rights set out in privacy rules and, where the service allows, choose a different arrangement. In a pay-per-use service, for example, someone who no longer wants a mileage-based formula can switch to a flat one, where distance data is no longer used to set the price.

‍What is the relationship between regulation and training in the use of AI?

‍
Regulation and training are closely linked. Regulation defines the perimeter you can move within, but training is what allows people to move inside that perimeter with greater awareness. It is important to avoid new technologies creating new forms of privilege for certain parts of the population. Everyone should have the chance to acquire the skills these tools require.

‍Why does training matter for personal data protection too?

‍
Because awareness is one of the first lines of defence. People need to be able to tell when a piece of data is being requested legitimately and when it is better withheld, and to recognise fraudulent communications and unusual requests for credentials or personal information. Training should also help users understand which data may be used to train AI tools and which settings they can change.

‍Is it more effective to protect consumers through stricter rules or through a better grasp of the technology?

‍
The two have to move together. Training and understanding the technology are a fundamental direction of travel, but regulation is necessary too, especially given the speed at which these tools are evolving. Competence has at least two dimensions: being able to use the tool properly, and being able to interpret and manage the answers the model produces. Using AI well therefore means knowing what the tool can do, understanding how it works and being able to assess its output critically.

‍What does it mean to have sufficient skills to use AI professionally?

‍
First of all, knowing how to use the tool in relation to the goal you want to reach. But also being able to interpret the answer produced and, where necessary, steer it. The model should be treated as an advanced tool, not as an entity that makes decisions on the user’s behalf. Training therefore has to cover both the technical use of the tool and the ability to interact with it and critically assess what it returns.

‍Should AI be used as the sole basis for a decision?

‍
No. In a work context in particular, you need to be able to interpret AI output in the light of other elements and other data, internal or external. The tool can support a decision, but the final assessment has to take account of the context and the information available. Critical thinking remains essential.

‍Can regulation keep pace with how fast AI is evolving?

‍
It is difficult, because these tools are evolving very quickly. At the same time, some rules are necessary, particularly in relation to the autonomy of these systems and their capacity to evolve and carry out increasingly complex tasks. Alongside regulation, developing public research also matters, so that knowledge of these technologies and their consequences does not remain solely in the hands of private operators.

‍When companies themselves ask for more rules, is that responsibility or strategy?

‍
Several readings of that behaviour came up in the discussion. One is that companies have perceived a pace of development that makes some form of regulation necessary. Another sees the call for rules as a possible strategic move: regulation could create barriers to entry and make it harder for new players to compete in model development. There is no definitive answer on the motivations. These are dynamics that will become clearer as the market evolves.

‍What is the main thing to watch for the future of AI?

‍
The speed of change is one of the defining features of the current phase. Adoption is growing fast and brings opportunities and risks at the same time. Early on, attention tended to fall on the opportunities: more information that could be analysed, speed, personalised services and new possible uses. Over time, a sharper awareness of the risks and a more grounded assessment of the actual benefits have begun to emerge.

‍Why does the consumer need to stay at the centre of AI’s evolution?

‍
Because it is people who will use these tools and, at least in civil applications, determine their success through adoption. Users therefore need to be put in a position to use AI as knowingly as possible, understanding both the benefits it can offer and the risks it carries. If the market evolves faster than consumers’ ability to understand its logic and advantages, a disconnect can open up between people and technology.

‍Where is the balance between opportunity and risk in using AI?

‍
There is no simply positive or negative answer. The data shows satisfaction and concern, opportunity and risk, at the same time. AI can increase the volume and speed of information analysed, make services more personalised and open up new opportunities. It also raises questions around privacy, data security, manipulation, training, and people’s capacity to understand and govern these tools. That is why awareness, training and regulation have to accompany technological change, keeping people at the centre.