**1. Introduction**

The world has recently witnessed the diffusion of the technological phenomenon of chatbots [1–4]. This phenomenon is simultaneously attracting and worrying public opinion, scholars and stakeholders. The attraction is due to the rapid diffusion, the easy accessibility, and the opportunities that chatbots, increasingly integrated with artificial intelligence, seems to offer. However, it is worrying that such rapid diffusion and easy accessibility have not been adequately accompanied by robust reflections on the impact they have on many domains of public life, from the social to the ethical and regulatory.

A chatbot can be defined as:


The introduction of artificial intelligence (AI) has inspired further stimulating scientific debates regarding its large-scale application, including in the world of healthcare [5]. An example of this is represented by the chatGPT tool [6], which has recently become widespread, rapidly attracting scientific attention to its potential and implications in its applications in social life [7] and in the health domain [8]. On the other hand, facing

the challenges of integrating technologies with AI is an open and current challenge with opportunities, challenges, and bottlenecks to overcome, affecting several domains [9,10], and also affecting learning processes and ethics [11].

The inclusion of chatbots is potentially disruptive in society, introducing opportunities but also important implications that need to be addressed on different domains.

The aim of this analysis is to examine chatbots in depth by mapping out their technological evolution, current usage, and potential applications within the health domain.

The sub-aims are:


#### **2. Methods**

The study was arranged into three points of view.

The *first point of view* traces the technological evolution of chatbots starting from the first pioneering experiences.

The *second point of view* reports the fields of application of the chatbots both from the point of view of categorization and of the sector of use, also giving space to the expectations of use and the expected benefits from a cross-domain point of view, also impacting the *health domain*.

The *third and main point* of view is that of the analysis of the state of use of chatbots in the health domain based on the scientific literature represented by systematic reviews. We decided to analyze the systematic reviews, the highest level of evidence in healthcare, because they are capable of providing a comprehensive evaluation of a particular topic by identifying and analyzing all the available primary research studies. This summary of evidence can help detect the principal patterns of interest and highlight areas where additional research is needed or where current research is insufficient to support clinical decisions, in this case, in the clinical domain.

The overview related to the first and second point of views was based on targeted searches on Google and Google Scholar.

The overview related to the third point of view followed a targeted search on PubMed by means of a properly settled composite key.

The overview, as a whole followed the ANDJ checklist, a standardized checklist for the structure of a narrative review.

This overview was carefully crafted with a consideration of five parameters (N1–N5) that have been evaluated on a scale ranging from one (minimum) to five (maximum). The parameters are as follows:

N1: Clarity of introduction and rationale for the review.

N2: Appropriateness of review design.

N3: Clear description of methods.

N4: Clear presentation of results.

N5: Justification of conclusions based on results.

N6: Full disclosure of potential conflicts of interest by authors.

These parameters have been thoughtfully selected to ensure the comprehensiveness and quality of this overview. All selected elements must receive a score of at least three on all parameters in order to be included.

#### **3. Results**

#### *3.1. An Overview of the Evolution of Technology*

The origin of chatbots [12] can probably be attributed to Alan Turing's 1950s vision of intelligent machines. Artificial intelligence, the basis of chatbots, has therefore, developed

these tools. We can summarize this evolution in brief [12]. The first chatbot named *ELIZA*, created in 1966, simulated a psychotherapist's function, repeating the users' sentences in an interrogative form. Its ability to communicate was limited; however, it can be considered a source of inspiration for further evolutions [13,14]. In 1972, *PARRY* was introduced. It acted as a patient with schizophrenia and defined its responses based on a system of assumptions and "emotional responses" [15]. AI was firstly used in the domain of the chatbots with the introduction of *Jabberwacky* in 1988. CleverScript was used in this system. It was a language based on spreadsheets, which was useful in the development of chatbots. This system was able to respond based on previous answers. It was limited in speed and number of users [16]. The term *CHATTERBOT* was introduced *in 1991*. It was an artificial player with a primary function of chatting [17]. *Dr. Sbaitso* appeared in 1992 [18]. It played the role of a psychologist seemingly without showing complications in its interactions with users [18]. *ALICE* was a further step forward in the world of chatbots. It used the artificial intelligence markup language. It performed better compared to ELIZA [19]. *SmarterChild* was introduced in *2001*. It was integrated with Messengers. This chatbot, for the first time, could help people with useful daily tasks using large databases with information related to movie times, sports, weather, and other information [20]. The chatbot using AI made another important step forward with the introduction of the smart personal voice assistants between 2010 and 2020. They were capable of understanding vocal commands and informative tasks. *Apple Siri, IBM Watson, Google Assistant, Microsoft Cortana, and Amazon Alexa* are the most popular voice assistants [21–27]. Early in 2016, a further evolution in AI technology radically changed the communicative interaction between users and manufactures. Social media platforms allowed developers to create chatbots that allowed the clients to complete specific tasks using their own messaging applications. At the end of 2016, the chatbots covered a wide range of applications ranging from entertainment to healthcare, including marketing, education, generalized support, cultural heritage, and much more. Moreover, the Internet of Things allowed new fields of application for chatbots; they played the role of connectors and mediators of "smart objects" [28]. See for example *Microsoft XiaoIce*, an AI-based chatbot with the role of satisfying the human need for sociability [29]. At the end of this process of evolution, the way of engaging in discussion with a chatbot was completely different from the *ELIZA* chatbot. Today, a chatbot is capable of sharing personal thoughts along with family drama events.
