Share

Articles - Archive

09/07/2019
Italian companies sceptical about artificial intelligence

Research Politecnico di Milano

 

In Italy, only 56% of all large companies have embarked on Artificial Intelligence projects, compared to about 70% in France and Germany.

 

La Stampa Tecnologia

The development of Artificial Intelligence - the branch of computer science that studies the development of hardware and software systems with skills we are normally accustomed to finding in humans, such as interaction with the environment, learning and adapting, thinking and planning - is still at an early stage, but is already making waves around the world, as evidenced by the growing interest of the academic world in the subject, offers by the largest Cloud service providers, the energy of 460 start-ups that, since 2016, have raised a total of 2.2 billion euro worldwide and, above all, the spread of AI solutions among large companies.

These are some of the results of a research performed by the Artificial Intelligence Observatory of the School of Management of the Politecnico di Milano, presented at the conference "Artificial Intelligence: perspectives from research to market". The research analyzed 721 companies and 469 cases of use of Artificial Intelligence, related to 337 international and Italian companies. In Italy, 56% of all large companies surveyed already have ongoing AI projects (against about 70% in France and Germany) highlighting the widespread attention towards the subject, even if we are still at an embryonic stage as far as the great opportunities offered by this technology: current projects are mainly oriented towards Intelligent Data Processing solutions (35% of cases) and Virtual Assistant/Chatbot (25%), while specific applications, such as supporting the internal processes of companies, are almost nonexistent. 

"At the moment, ready-to-use solutions are limited and to reach a level of performance similar to or higher than human, requires a lengthy preparatory phase for the infrastructures, information heritage, skills and culture, as well as constant work-in-progress adjustments in machine learning. AI projects need perseverance and a result-oriented approach, rewarded not only with performance, but also, and more importantly, with a competitive advantage in terms of know-how”.

In detail, the research, after analyzing 469 cases, revealed that only 38% of AI projects identified globally are actually up and running (popular among every-day users and subjected to a continuous well-structured development process). "One in five, 21%, is being implemented on a large-scale - notes Alessandro Piva, director of the Artificial Intelligence Observatory - involving a significant share of processes and users. On the other hand, a similar proportion is still in its pilot phase, during which the first results are measured and problems identified, while 20% is still only on paper, although funds have been allocated to explore projects in well-defined fields of application”.

Compared to what is happening abroad, where experimental applications are also being carried out, Italian companies seem to prefer more established settings: 52% of the solutions identified are already up and running (compared to 38% at international level), while 48% are still either on paper, pilot project or in their implementation phase. Most AI projects in Italy are finalized in fields such as banking, finance and insurance (17%), automotive (17%), energy (13%), logistics (10%) and telco (10%). "Potentially, Artificial Intelligence – according to Nicola Gatti, Giovanni Miragliotta and Alessandro Piva, directors of the Artificial Intelligence Observatory - knows no application boundaries and will progressively affect the economic and social fabric of each country. Its rapid expansion in several areas, however, will not be homogeneous, but will largely depend on factors such as technology and know-how. Italian enterprises are paying great attention to this key aspect so as not to miss the opportunity to improve their competitiveness".

"Reaping the potential benefits, however, depends on fully understanding the range of available solutions - they continue - and then work on organizational and management processes as well as enhancing personal skills. The research shows that any project of Artificial Intelligence, when starting up, requires large investments and not only in economic terms”.

"We are only at the beginning - says Nicola Gatti, director of the Artificial Intelligence Observatory - of a long journey of understanding and diffusing the potential benefits of Artificial Intelligence, which will define its boundaries of application and the degree of intelligence in any solution. From self-driving cars to household appliances that learn and adapt to the lifestyle and needs of families, from personal assistants who advise us on spending decisions to the role of robots in human care-taking, all daily activities will be rethought in the light of the capabilities of “intelligent” machines. The speed of this transformation will largely depend on the existence of consolidated technological solutions, the ability to manage a delicate change in organizations and the balance between the inherent value of such innovations and the costs involved in making intelligent products and processes”.

Looking at the sectors, banking, finance and insurance companies appear the most active in introducing Artificial Intelligence solutions at an international level, which collects 21% of existing solutions, driven by the opportunity to get to know their customers in greater detail and ensure targeted services and management support. A little behind we find the automotive sector (12%), driven by large investments aimed at developing self-driving vehicles. Further behind, with percentages between 6% and 8%, are hi-tech, retail and telco companies, interested in offering a more flexible and personalised service. The other sectors are rather marginal but still active, testifying to the high pervasiveness of artificial intelligence, with percentages between 3% and 5%. On the other hand, analyzing the company functions, processes dedicated to customer relations (marketing, sales and customer service), collect 40% of the solutions, while in all the other internal processes (Operations, Hr, Research, Development and Finance) reach a further 40% leaving the remaining 20% to deal with the products offered, with the aim of increasing their performance and user experience. On the other hand, the emphasis on efficiency boosting solutions and labour costs reduction appear to be playing a minor role.

Looking at the potential repercussion on employment, there are a number of concerns about the negative impact of artificial intelligence. The Observatory investigated the issue, analyzing the evidence of international and national publications (scientific or opinion leaders) on the impact of AI on employment as well as the cognitive, psychological and social implications of replacing human labour with machines. Three encouraging signs emerge, which seem to contradict the gloomiest forecasts. Although we are in a transitional phase and ready-to-use solutions are rare, the analysis shows that the demand for labour in AI projects has increased, not decreased. Moreover, AI solutions are now used more as an external competitive lever to improve services and quality than as tools to increase internal efficiency and productivity. Finally, companies appear to be aware of the sensitivity of the issue, carefully selecting the projects, considering both the expected benefits and the internal and external acceptability of the innovation.

"The encouraging results - comments Giovanni Miragliotta, Director of the Artificial Intelligence Observatory – teach us to adopt a non-alarmist tone, being careful, though, not to run the risk of looking at things through rose-coloured lenses. It would be wrong, in fact, to embrace Artificial Intelligence with uncritical and superficial enthusiasm: we need in-depth analysis, sense of responsibility, and strategic projects that meet the concerns and expectations of the workforce".

back to archive