CHATBOTS AND BIG DATA: How conversational agents collect, analyse and exploit big data.

The last few years have been characterised by innovation and revolution. A particular aspect is worth to consider is the so-called Digital Era. The Digital Era is a great revolution that changed social, economic and legal aspects of our society. The birth of the Digital Era involves three significant changes. The first innovation concerns the representation of new forms of expression, such as sounds, images and texts, in binary system thanks to the new technologies. The second one is about elaboration of the information thanks to new automatic tools, like computers. The third, and last, big change is about communication made possible because, thanks to computers, the elaborated bits can transfer information from one part of the world to another in a very short time. In this scenario we can place Big Data: «a large, diverse sets of information that grow at ever-increasing rates. It is characterised by the large volume of information, the velocity at which it is created and collected, and the variety or scope of the data points being covered». This study focuses on the analysis of specific artificial intelligence software capable of structuring a conversation with the user: they are called chatbots. In this process chatbots use big data to reach their aim: responding as coherently as possible to the question asked by the users, providing the requested information, showing the content on which it has been questioned or guiding the user through the path of discovery, interest, in-depth analysis and action that the developer has defined. This thesis is divided into five chapters: the first chapter introduces the meaning of chatbots, their importance and the first attempt to create a conversational agent. The second chapter illustrates how chatbots work, depending on their type. The third one, more consistent, analyses the relationship between chatbots and big data; in detail, we will see how they collect, analyse, exploit and improve big data. In the second part, corresponding to the fourth and fifth chapters, the problem of unsupervised learning and the issue of privacy and personal data are illustrated.