"Hello, my name is TikTok. I'm a chatbot in the Chatbot project, providing advice on gaming addiction problems," TikTok greets users immediately after they scan the QR code or add the LINE account @healthygamer. When users request advice via the chatbot, they'll be presented with many options in three main sections. For example, "How to know if your child is addicted to games." Clicking on this section will lead to options for taking quizzes on gaming addiction, online media addiction, parenting, and an internet gaming addiction assessment – both for individuals and parents – to measure the severity of their symptoms and determine if treatment is needed. There are also frequently asked questions, videos, and infographics providing information about gaming addiction, warning signs, dangers, dos and don'ts, and proper prevention.
The "Chatbot for Consulting on Gaming Addiction" has been available for free for over a year. It's the brainchild of Assoc. Prof. Dr. Woraphat Rat-arpa and his team from the Department of Psychiatry, Faculty of Medicine, Siriraj Hospital, along with academics who have been researching and developing media to address children's gaming addiction for almost 20 years. Their goal was to create a tool that allows parents access to accurate and scientifically-based initial advice and information without having to travel long distances or wait in long queues to see specialists. They presented this idea to the Safe and Creative Media Development Fund and received collaborative funding in 2021 to develop the AI chatbot, "Nong TikTok," for addressing gaming addiction. The development and testing process, including content and communication, took almost a year, requiring approval from experts and qualified personnel before the data was input into the chatbot and undergoing real-world testing.
The development team's goal is for the chatbot to understand language and answer questions as if it were conversing with a person. This is quite difficult, so when the bot responds to a user, the team reviews the answer to see if it's appropriate. If it is, it learns to use that type of response; if not, it changes the wording to learn how to answer better. Furthermore, frequently asked questions and commonly used words will be linked to the targeted responses more quickly.
“There are several challenges and obstacles. The first limitation we're trying to overcome is the accuracy and the bot's ability to interpret natural language to understand human language. This includes understanding the text users type and continuously engaging in user conversations. The bot finds this difficult, for example, sometimes encountering conversations with overly long sentences or words that haven't been learned before. It becomes challenging for the bot to understand and answer the questions.”
During the initial development phase, real psychologists were used to answer questions instead of the bot. Users could choose to chat directly with a psychologist, and if a user had a serious problem, they would be recommended treatment. Those with less serious problems could receive consultation through the chatbot until completion. The team would then input the conversations and messages into the system for the bot to learn from, and as it learned more, errors would decrease.
"In the bot's operation, continuous improvement is still needed because it's very important. If it can't communicate effectively, users won't want to interact, especially with sets of questions that relate to human behavior and are very diverse, as well as the nuances of word and sentence usage in Thai. Furthermore, most people prefer face-to-face interaction. Therefore, chatbots may not be very popular yet for functions requiring extensive problem-solving."
The feedback received included both praise and suggestions. Most said that the chatbot was very useful but should be developed to understand human language, the complexity of sentences, and respond more directly to questions.
“These feedbacks, whether positive or indicating areas for improvement, I feel help us better understand our users—what they really want, how they want to use it, and what their needs are,” Asst. Prof. Dr. Woraphat concluded with a joyful tone, pleased that the team was at least on the right track to some extent, that people recognized the value and found these things beneficial, and that they could help solve the problem of children addicted to games in society.
#MediaFund #ChatbotForGamingAddictionConsultation
#SharingStories #ArticleSharingStories
#Projects by Media Fund Grant Recipients
#Creative media for social change
#Fund for the Development of Safe and Creative Media
