In the midst of contrasting opinions, fideistic enthusiasm, fear for employment, risks of discrimination and utopian predictions, Artificial Intelligence (AI), a hot topic in recent years, is experiencing a new dawn. Although we are talking about it regularly, a veil of confusion often hovers over it and it isn’t always clear which technologies we are referring to when we mention AI.
First of all, a distinction needs to be made between Artificial General Intelligence and Weak Artificial Intelligence. The first intends to create a kind of artificial “brain” that is able to reason and fully emulate human beings in all their skills and abilities. The second applies the learning logics and mechanisms of AI engines to specific areas of knowledge in order to study or solve problems where technology is able to go way beyond human cognitive limits. To give an example, AI systems may find space in areas of medical application, since algorithms, if trained correctly, are able to identify a melanoma 99% of the time (compared with 85% in “human” specialists).
The name Artificial Intelligence has its roots far back in time, specifically in 1956. It was created by several intellectuals who were working on computational technologies that they compared with one another during a conference in New Hampshire. In the decades that followed, the topic fell into the background because a fairly unsurmountable obstacle emerged: these neural networks needed to be trained with large amounts of data that, at the same time, required considerable computational skills to collect and elaborate them.
With the recent technological progress brought about by the digital revolution, including the increase in computing power (ever more available with as-a-service models in the cloud), the availability of large quantities of data, and increasingly sophisticated tools for their analysis, Artificial Intelligence has seen an “acceleration” to all intents and purposes.
Within the Digital Innovation Observatories at the Polytechnic University of Milan, we have developed the following definition: “Artificial Intelligence is the branch of computer science that studies the development of hardware and software systems equipped with capabilities that are typically found in human beings and that are able to independently pursue a defined end, making decisions that, beforehand, were usually entrusted to human beings.”
The learning and natural language recognition algorithms that AI technologies are equipped with make them able to interact with external environments, giving answers, including physical ones, to the interlocutors that approach them. In the case of autonomous systems, they make decisions without human interference. In any case, AI is a “moving target” that continues to raise the bar. In the future, we will see more and more tasks that were originally carried out by humans in the “hands” of machines.
AI applications: a Polytechnic study
Recently, the Digital Innovation Observatories at the Polytechnic University of Milan carried out a survey to understand what the application cases of AI are. The study focused on the analysis of large international companies with a turnover of over $15 billion (the level is very different compared with Italian companies) and took into consideration 469 initiatives in 337 companies, evaluated according to certain parameters. What emerged were eight classes of solutions that were studied both from the point of view of the diffusion of the relative applications and from the point of view of their applicability in specific industries.
Solution Classes and their diffusion
- Autonomous Vehicles (7%), or rather, self-driving vehicles dedicated to the transportation of people, animals or things, either on the road (vehicles), via sea, lake or river navigation (floating craft), or, finally, in flight through the atmosphere or in space (aircraft), capable of perceiving the external environment and identifying the correct maneuvers in order to adapt to it.
Fears of negative implications are not unfounded. Just think about the drones for war purposes in military environments and what could happen if they ended up in the wrong hands.
- Autonomous Robots (4%), or rather, robots capable of moving themselves or some of their parts (arms, etc.), manipulating objects and performing actions of various kinds without human intervention, drawing information from the surrounding environment and adapting to unforeseen or unencoded events. This class ranges from toy robots and gadgets to extremely complex applications that cover everything from the manufacturing of the intelligent automation of some production processes to robots that simulate movement (for example, those modeled on animals created by Boston Dynamics).
- Intelligent Objects (7%), or rather, objects that are able to perform actions and make decisions without requiring human intervention, interacting with the surrounding environment through the use of sensors (thermometers, camcorders, etc.), actuators (opening/closing windows/doors, activation of household appliances and systems, etc.) learning from the habits or actions of the people who interact with them. This ranges from the suitcase that follows its owner on their journey to smart video cameras that are able to identify how people are moving outside. China is one of the most interesting markets in this field but also with negative implications, given that they could be used to control people.
- Virtual Assistants/Chatbots (25%), or rather, software agents that are able to perform actions and/or provide services to a human interlocutor based on commands and requests recognized via a natural language interaction (written or spoken). It is one of the most interesting areas of application on a business level as they can be used to interact with customers both in after-sales support and during the sales process.
- Recommendation (10%), or rather, solutions aimed at guiding user preferences, interests and, more generally, decisions, based on information provided by them directly or indirectly, with outputs consisting of personalized recommendations that can be positioned at different points along the customer journey or, more generally, during the decision-making process. In particular, these technologies accompany product selection and upselling.
- Image Processing (8%), or rather, image analysis solutions, individual or sequential (video), aimed at recognizing people, animals and things present within the image itself, biometric recognition (e.g., facial, iris), and, generally, extracting information from images. Just like language processing, these technologies are often used to support the development of other applications.
- Language Processing (4%), or rather, language elaboration solutions whose purposes can range from the understanding of content and translation to the independent creation of texts starting out from data or documents provided in input.
- Intelligent Data Processing (35%), or rather, solutions that use AI algorithms on structured and unstructured data for purposes relating to the extraction of information present in the data (predictive analysis, pattern discovery, etc.).
In the sample of companies surveyed, it was found that the most advanced initiatives concern virtual assistants/chatbots and recommendation, language and intelligent data processing, while, as far as interest in the implementation of AI is concerned, it is high in the banking, insurance and automotive industries in particular.
- The high pervasiveness of AI, which can be applied to all industrial sectors (with a high number of initiatives in banking, insurance and automotive industries), can certainly be observed.
- From the study, a completely “virgin” Italian market emerges, which is also a late developer when compared with the big European countries (France, Germany, the UK, etc.).
- Furthermore, it can be observed that, at the moment, chatbots/virtual assistants are a quick-win project, or rather, an easy earner, but they are not always approached in the right way.
Artificial Intelligence and Content Marketing: a potential synergy?
Artificial Intelligence offers interesting opportunities in many industries, which certainly include that of content marketing too. On the one hand lies the possibility to generate simple content automatically, allowing you to have a starting point from which to enrich content with information of added value.
On the other hand, it offers new tools to make interaction with the end user more functional and effective. We are talking about, for example, the possibility of having algorithms that are able to find out what the user’s tastes are in an incredibly detailed way, or the ability to interact with completely different logics based on natural language and chatbots.
At the end of the day, Artificial Intelligence therefore represents an enabler to enrich interaction with customers, giving new stimuli and information to those in the industry whose job it is to develop an offer that is always new and keeps up with the times.
Among the new frontiers of Content Marketing is Content Intelligence, a term used to indicate AI systems and software that measure the performance of content from the use that users make of it.