Diabetes and conversational agents: the AIDA project case study

Abstract

One of the key aspects in the process of caring for people with diabetes is Therapeutic Education (TE). TE is a teaching process for training patients so that they can self-manage their care plan. Alongside traditional methods of providing educational content, there are now alternative forms of delivery thanks to the implementation of advanced Information Technologies systems such as conversational agents (CAs). In this context, we present the AIDA project: an ensemble of two different CAs intended to provide a TE tool for people with diabetes. The Artificial Intelligence Diabetes Assistant (AIDA) consists of a text-based chatbot and a speech-based dialog system. Their content has been created and validated by a scientific board. AIDA Chatbot-the text-based agent-provides a broad spectrum of information about diabetes, while AIDA Cookbot-the voice-based agent-presents recipes compliant with a diabetic patient's diet. We provide a thorough description of the development process for both agents, the technology employed and their usage by the general public. AIDA Chatbot and AIDA Cookbot are freely available and they represent the first example of conversational agents in Italian to support diabetes patients, clinicians and caregivers.

Supplementary information: The online version contains supplementary material available at 10.1007/s44163-021-00005-1.

Fig. 1

Participants and goals of the two interactive workshops

Fig. 2

Some example dialogues. In Example 1, the user decides to consult AIDA by interacting with the content menu. In Example 2, the user asks a very specific question and AIDA tries to match it to the closest one in its KB. In Example 3, the user asks a series of questions related to the same topic. AIDA keeps track of the current topic in order to answer appropriately. The dialogues, originally in Italian, have been translated into English for the purpose of this article

Fig. 3

In Example 1, the user interacts with AIDA Cookbot for the first time and the agent asks for any allergy, in order to memorize it. Users can ask a Diet-related question or for a recipe. In Example 2, the user returns for a second interaction and expresses a preference for a recipe. AIDA guides the user towards something more specific. In Example 3, the user asks directly for a recipe and AIDA immediately proposes something related to the request

Fig. 4

AIDA Chatbot architecture. The figure shows the interlacing between the touch points and the backend modules: the dialogue manager, the NLU and NLG modules, the reasoning ones (AIDA relies on a machine learning system as well as a rule-based engine). The communication with the touch points is established via API rest, while the contents of the agents themselves are managed in a separate database, through a specific content manager graphical interface

Fig. 5

AIDA Cookbot’s architecture. The direction of the arrows symbolizes the flow of the information: from the Alexa endpoint towards one of the two models, down to the understanding engines. Once the proper answer has been crafted, it is passed up again through the up-pointing arrows

Fig. 6

The initial states of AIDA Cookbot’s interaction model