Ai agents, privacy and ethics decrease trust

People are quick to talk about Ai agents, but introducing them into organisations seems more complicated than expected. The strong initial enthusiasm has given way to caution, especially over two issues, privacy and ethics, which have caused trust to drop dramatically, from 43% of organisations to 27%, despite the fact that Ai agents would seem to be 'colleagues' capable of developing a very high potential in organisations. In fact, a report by the Capgemini Research Institute, 'Rise of agentic AI: How trust is the key to human-AI collaboration', estimates that agent-based Ai will be able to generate an economic value of up to 450 billion dollars by 2028 globally. Despite this, however, on the one hand the number of companies that say they trust is falling, and on the other hand two out of five executives globally believe that the risks outweigh the benefits at this stage, according to the report, which was carried out on 1,500 executives from organisations with annual revenues of more than $1 billion, in 14 countries, including Italy, and belonging to 13 different sectors.

In fact, the very rapid evolution of new tools related to artificial intelligence shows the need for organisations to adapt to them in a way that is sustainable for everyone. And if there is an important percentage among those that use AI agents in a support function, it is a different matter when it comes to autonomous use. Only 40 per cent of companies say they trust AI agents to manage tasks and processes autonomously, while the majority do not trust the technology completely.

Italian organisations, in line with the global trend, have less trust in artificial intelligence agents than in 2024. About 60 per cent of Italian organisations are concerned about privacy issues, security risks and the risk of unwanted bias related to the use of Agentic Ai. Over the next 12 months, Italian organisations expect collaboration with Ai to result in an increase in team member skills, while with a three-year perspective, the same organisations envision Ai agents as autonomous members within human-supervised teams.

The Italian cross-section of data shows a different approach in the short and medium term and an evolution of the functions involved. Thus, in the next 12 months, among the corporate functions where organisations expect Ai's agents to manage at least one process or sub-process per day, the functions of Sales, IT, Marketing&Communication and Finance stand out. When the time span is extended to three years, the list gets longer and operations are also among the functions that Italian organisations expect to be most impacted by Agentic Ai.

For Monia Ferrari, Managing Director of Capgemini in Italy, 'the economic potential of AI agents is enormous, but translating it into reality requires more than just technology: it requires a strategic and structured transformation involving people, processes and systems. To be successful, organisations must focus on results, rethinking their processes with an 'AI-first' vision. At the heart of this transformation is the need to build trust, developing AI in an ethical and safe way from the very beginning. Organisational structures also need to be redesigned to foster an effective synergy between humans and AI, creating the conditions to enhance human judgement and achieve better business results."