Abstract:

Social networks have become the communication media par excellence among people. Through them, friends and family interact, in addition to consulting sources of relevant information. However, currently, it is difficult to differentiate between true and false information. In an effort for trying to fight against fake news and disinformation, two platforms have been used for the development of educational Chatbots created with the purpose of generating questions in users that lead them to ask themselves questions such as: What kind of information do I receive? What are the reliable sources of information? What emotions does the information I receive provoke in me? Is this information true or false?

Objective:
The objective of this work was to analyze two platforms for the creation of educative Chatbots.

Method:
This is a descriptive study that compares two platforms to create chatbots from dimensions of interest to teachers.

Instruments:
Platforms for creating chatbots: sendpulse.com and dialogflow.cloud.google.com.

Procedure:
Due to the need for creating chatbots from the point of view of teachers, not programmers, the following dimensions were used for the evaluation of the two platforms: type of questions and answers, ease of use, social networks where they can be used, number of messages per day, cost of the platform, the possibility of replicating the flow of conversation, and access to the conversations generated by users.
In 2022, the sendpulse platform was used to create a chatbot, tested with a pilot group.
In the first half of 2023, a chatbot was created, as well, on Google's Dialogflow platform and tested with a pilot group.

Result:
As a result of this analysis, two different types of chatbots were obtained.
By employing sendpulse, it was found that responses can be closed: yes, no, or multiple choice. This includes open-ended responses that direct to the next step regardless of the response’s content. It can be integrated into Facebook Messenger and Telegram, while data can be downloaded from CSV files for analysis. The paid version offers more qualities than the free version.
On the other hand, with the Chatbot developed with Dialogflow, a more open dialogue was obtained, it does not include closed responses, as it is based on Natural Language Processing, so it takes keywords and issues a programmed response, but keywords cannot be reused, unless they are configured with a "context". It comprises Facebook Messenger and may include WhatsApp from the triangulation with a cell phone linked to an autoresponder. You don't need to pay to work with the free version (Dialogflow ES).

Conclusions:
Both chatbots have proven to be useful and functional; however, there must be a balance, to be found, between closed and open responses that allow probing into the user's way of thinking. This is important for teachers since not all of them have programming skills, so it is essential that chatbot platforms have user-friendly features. The conclusion is that it is important for teachers to continue with research to generate increasingly sophisticated educational chatbots.

Keywords:

Chatbot, dialog flow, education, social media.