DOI: https://doi.org/10.21203/rs.3.rs-2565141/v1
Background: Artificial intelligence (AI) Chatbots are computer programs that simulate human conversation and use artificial intelligence including machine learning and natural language processing to interact with users via natural language. With the outbreak of the COVID-19 pandemic, the use of digital health technologies such as chatbot has accelerated. This study aimed to investigate the application of AI chatbots in combating COVID-19 pandemic and to explore their features.
Methods: We reviewed of literature on health chatbots during the COVID-19 pandemic. PubMed, Scopus, Web of Science and Google Scholar were searched by using related keywords such as "chatbot", “conversational agent” and “Artificial intelligence”. To select the relevant articles, we conducted title, abstract, and full-text screening based on inclusion/exclusion criteria. Chatbots, their applications and design features were extracted from selected articles.
Results: Out of 673 articles initially identified, 17 articles were eligible for inclusion. We categorized selected AI chatbots based on their roles, applications and design characteristics. 70% of chatbots had preventive role. Our review identified 8 key applications of the AI chatbots during the COVID-19 pandemic that includes: 1) information dissemination and education 2) self-assessment and screening 3) connect to health centers 4) combating misinformation and fake news 5) patients tracking and service delivery 6) mental health 7) monitoring exposure 8) vaccine information and scheduler. AI chatbots were deployed on various platforms including mobile apps, web and social media. Mobile-based chatbots were the most frequent. All of chatbots use NLU methods to understand natural language input and act on the user’s request. More than 50% of AI chatbots were used NLU platforms including Google Dialogflow, Rasa framework and IBM Watson.
Conclusion: The AI chatbots can play an effective role to combat COVID-19 pandemic. Increasing people's awareness, optimal use of health resources, reducing unnecessary encounters are some advantages of using AI chatbots during COVID-19 outbreak. Using NLU platforms can be a suitable solution in the development of AI chatbots in healthcare domain. With advancement of Artificial intelligence field, it seems that AI chatbots will mark a bright future in healthcare specially in public health, chronic diseases management and mental health.
With the spread of the COVID-19 pandemic around the world, significant demands were imposed on health care systems, and the effective use of available resources and capacity expansion became a vital matter[1]. The various restrictions such as social distancing to slow and prevent the transmission of the virus created major challenges on the provision of the traditional and face-to-face services[2]. Furthermore, the rapid growth and update of information about the virus and the fast spread of misinformation and fake news created confusion and anxiety[3, 4].
The use of chatbots and other digital health technologies are the responses to these issues health care challenges[5, 6]. A chatbot, also known as conversational agent, is a computer program that simulates human conversation and interacts with users in natural language through voice commands or text chats or both[7, 8]. There are two main types of chatbots, Rule-based and artificial intelligence-based (AI) chatbots[9]. Rule-based chatbots communicate using predefined answers. These chatbots provide answers based on a set of if-then rules and don't understand the context of the conversation. They provide matching answers only when users use a controlled and predefined keyword or a command (like Telegram bots). However, if anything outside the agent's scope is presented, like a different spelling or dialect, it might fail to match that question with an answer. AI chatbots use artificial intelligence including machine learning and natural language processing (NLP) to interact with users via natural language[10]. Machine Learning allows chatbots to identify patterns in user input, make decisions, and learn from past conversations. NLP helps chatbots to extract structured data from unstructured language input and enable them to replicate that behavior. NLP lets them understand the context of the conversation even if a person makes a spelling mistake. Therefore, the users can freely speak or write to the chatbot[11]. AI chatbots can be better than their rule-based types[12, 13].
Recently the use of chatbots has increased across industries and business functions such as customer service, sales and marketing[14]. In the healthcare, chatbots have been used for various purposes such as mental health, maternal health, physical activities and nutrition [15–17]. The use of chatbots can improve access to health information and services to millions of people at once[18–20]. Many organizations and governments developed health chatbots to provide accurate information about COVID-19[18, 21]. For example, the World Health Organization (WHO) developed a chatbot to fight COVID-19 on the WhatsApp platform, which disseminates news and new findings related to the COVID-19 and answers to the users’ questions on protection and care against the Coronavirus[22].
This study aimed to review the recent literature on AI chatbots related to COVID-19, identify and describe their applications for fighting the coronavirus, and investigating the design features of these chatbots. This review will provide insights into current AI chatbots research in healthcare, applications and technical aspects that can be effective in the design and use of this initiative in healthcare and management of pandemics such as of COVID-19 that may occur in the future.
We conducted a scoping review according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) principles[23]. Literature searches were performed in the electronic databases of PubMed, Scopus،Web of Science and Google Scholar. For Google Scholar, we only used the first 150 results because Google Scholar returns the most relevant results for the search term first. The search was conducted between November 2019 to October, 2022. Articles were searched in databases by combining related terms such as "chatbot", “conversational agent” and “COVID-19”. Search terms were categorized into 2 groups. To combine terms, we used the OR operator for inside each group and the AND operator for between groups (Additional file 1).
To select the relevant articles, we used the following inclusion criteria: original articles, full text articles and studies that used AI chatbots for COVID-19. We excluded studies that a) were not focused on the development or use of a chatbot, b) chatbots were not related to health domain c) chatbots were not used for fighting COVID-19 d) were used Rule-based chatbots. Non-English and review articles were also excluded.
Duplicate studies were first identified and eliminated, then in order to minimize bias in the selection of the article, 2 authors independently (M.A and A.M) conducted title, abstract, and full-text screening. Disagreements between the reviewers about the selected articles were resolved by a third author (R.N).
Two investigators independently extracted data from studies (M.A and A.M). The data extracted included general study information (such as author’s name and publication year), description of chatbot (chatbot name, aim and language), chatbot applications and technical characteristics (such as platform, AI techniques, and user-chatbot interaction). Extracted data were recorded in a spreadsheet. Any disagreements about the data extracted were resolved by the third reviewer (M.H). The data extraction form is presented in additional file 2.
Figure 1 illustrates the studies selection process. A total of 673 articles were identified from electronic database searching. After removing 313 duplicate articles, the abstracts and titles of articles were screened and 271 studies were excluded. After the full text reviewing phase and removing 72 studies, 17 article were selected finally.
The most common countries of publications were the United States (n = 4), India (n = 4) and Canada (n = 2). Australia, Spain, Germany, Hong Kong, UAE, New Zealand, and Austria, each had published one study each. The Studies were published between 2020 and 2022 (7 articles in 2022, 6 articles in 2021 and 4 articles in 2020).
The total number of AI chatbots reviewed was 17 (Additional file 3). 14 chatbots (88.2%) conversed in English language. 4 chatbots were bilingual and in addition to English, supported French, Italian, Vietnamese and Indian languages. 3 chatbots were also multi-language. Arogya setu and COVIBOT that were implemented in India supported more than 10 common languages in India[7, 24].
13 of 17 chatbots were developed for general population and 4 cases for specific population. Most AI chatbots with general audience were developed by international organizations (such as WHO), governments and large companies (such as IBM and Amazon). For instance CoronaGO and Arogya setu were designed by Indian government to improve care delivery and help to its people during COVID-19 situation[7, 25]. AI chatbots with specific audience were developed for young people (2 cases) [26, 27], elder people (1 cases) [28], and health care providers and their families (1 case) [29].
We categorized AI chatbots based on a) their aims which reflects the chatbot role(s) in response to COVID-19, b) their applications, and c) their design characteristics which include platforms, AI techniques and chatbot-user interaction design.
According to the study by Jovanovic and et al [30], We classified The AI chatbots based on the chatbot role in healthcare provision into 3 groups, preventive, diagnostic and therapeutic (Table 1). 12 cases were preventive chatbots. The purpose of these chatbots were to inform and educate users about how to take care and protect themselves against COVID-19. Diagnostic chatbots (9 cases) were developed to checking user’s symptoms and self-assessment. These chatbots help to disease screening and support to the COVID-19 diagnoses. Therapeutic chatbots (3 cases) supported and assisted to therapeutic services provision to patients. For example, support to providing tele-consultations and home care for COVID-19 patients[24]. 7 of 17 chatbots had multiple roles in health service delivery that were 5 cases preventive and diagnostic, and 2 cases diagnostic and therapeutic. Ten chatbots were single purpose that were 7 preventive, 2 diagnostic and 1 therapeutic.
|
Chatbot Name |
Chatbot Role |
Chatbot Application |
||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
|
Preventive |
Diagnostic |
Therapeutic |
Information dissemination |
Self-assessment and screening |
Exposure monitoring |
patients Tracking and service delivery |
combating misinformation and fake news |
Mental health |
Vaccination information and scheduling |
Connect to Health care center |
||
|
Arogya setu[7] |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
|||||
|
Alexa[35] |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
|||||
|
Symptoma[31] |
✓ |
✓ |
||||||||||
|
CovidBot[28] |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
||||||
|
Aroha[27] |
✓ |
✓ |
✓ |
✓ |
||||||||
|
Chloe[32] |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
||||||
|
VWise[36] |
✓ |
✓ |
✓ |
✓ |
||||||||
|
CoronaGO[25] |
✓ |
✓ |
✓ |
✓ |
✓ |
|||||||
|
Cory COVID-Bot[47] |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
||||||
|
COVIBOT[24] |
✓ |
✓ |
✓ |
✓ |
✓ |
|||||||
|
MIRA[29] |
✓ |
✓ |
✓ |
✓ |
||||||||
|
Vac Chat, Fact Check[26] |
✓ |
✓ |
✓ |
✓ |
||||||||
|
Watson Assistant[48] |
✓ |
✓ |
||||||||||
|
WashKaro[37] |
✓ |
✓ |
✓ |
✓ |
✓ |
|||||||
|
Woebot[49] |
✓ |
✓ |
||||||||||
|
Ada Health[33] |
✓ |
✓ |
✓ |
|||||||||
|
SanIA[34] |
✓ |
✓ |
✓ |
✓ |
✓ |
✓ |
||||||
|
N (%) = |
12(70.58%) |
9(52.94%) |
3(17.64%) |
13(76.42%) |
9(52.94%) |
4(23.53%) |
5(29.41%) |
7(41.17%) |
5(29.41%) |
3(17.64%) |
8(47.05%) |
|
The reviewed chatbots provided different services to users. We classified the chatbots applications into 8 groups (Table 1), which include 1) Information dissemination and education, 2) Self-assessment and screening, 3) Connect to health centers, 4) Combating misinformation and fake news, 5) Patients Tracking and providing services, 6) Mental health, 7) Exposure monitoring, and 8) Vaccine information and scheduling.
1) Information dissemination and education
The most common application of AI chatbots was Dissemination of information and provision of education related to COVID-19 (13 cases). These chatbots provide the various information such as symptoms, diagnosis of disease, stages of the disease, virus transmission, and availability of healthcare services and how to access them. These chatbots also provide various trainings related to preventive and care practice (such as cleaning hands, wearing face masks).
2) Self-assessment and screening
Nine AI chatbots were utilized as triage and assessment of users during the COVID-19 pandemic. These chatbots ask a series question in order to assess the possibility of COVID-19 infection without going to medical centers. Self-screening chatbots check the COVID-19 symptoms based on the suggested guidelines by governments and approved sources (such as CDC or WHO) and make recommendations if needed. For instance, the Symptoma assess the possibility of a person contracting coronavirus by receiving information such as disease symptoms, age and gender. The results of a study have shown that Symptoma was able to correctly diagnose users with COVID-19 in 96.32% of cases [31].
3) Connect to health centers
Users supporting to communicate with health centers and healthcare providers was another use cases of AI chatbots(n = 8). Some chatbots, in addition to evaluating and triaging people, were able to identify the user’s location by GPS technology. These chatbots provide information about nearest health centers to the users and make an appointment for further investigation[7, 28, 32]. Some AI chatbots were able to communicate online with doctors and health centers so that, the users can benefit from remote diagnostic and consulting services[7, 33]. SanIA, was developed in Spain, created a secure communication channel between the citizen and the health system during the COVID-19 outbreak, to maintain continuity of care by providing advice, including psychological assessment, to patients whenever they want it 24/7[34].
4) Combating misinformation and fake news
7 of 17 AI chatbots were able to combat misinformation and fake news about COVID-19. These chatbots provide users with reliable and accurate information such as up to date statistical information, best practices and health protocols[28, 35]. These chatbots monitor and identify misinformation and rumors about COVID-19 and prevent its further spread by alerting users.
5) Patients tracking and service delivery
In 5 cases of AI chatbots, it was possible to provide remote care services to people with COVID-19. These chatbots were able to continuously and daily track the patient’s physical condition and vital signs. They also answer patients’ questions 24/7 during mandated self-isolation periods and if necessary, help to communicate with health care centers and receive advice and remote care services[24, 32]. For example, COVIBOT helps the patient to take proper medication and remote consultation[24].
6) Mental health
Supporting people’s mental health during the pandemic and providing necessary services was another application of AI chatbots(n = 5). People’s mental and emotional problems such as stress, anxiety and depression were one of the problems that existed during the COVID-19 pandemic. 5 of AI chatbots were able to monitoring and Supporting people’s mental health during the pandemic. These chatbots by providing various mental health services to users, specially to patients in home quarantine, help to reduce the psychological effects of COVID-19 such as anxiety, distress and depression. For example, the Aroha was an AI chatbot that provided services related to the management of anxiety and depression caused by COVID-19 to New Zealand citizens[27].
7) Exposure monitoring
Four chatbots were deployed for monitoring exposure to the Coronavirus and providing notifications. These chatbots provide alerts and necessary information to users by assessing the status of disease spread, as well as identifying high-risk and infected locations. The Arogya Setu works on Bluetooth-based and GPS technologies and tries to determine risk based on the user’s location. It also keeps the user informed in case he/she has crossed paths with the positive COVID-19 case within 6 feet proximity[7].
8) Vaccine information and scheduling
Three chatbots were designed for providing services about COVID-19 vaccines. These chatbots provide accurate and up to date information about COVID-19 vaccines and informing users with their advantages. In addition, they provide information about vaccination centers and support to make an appointment for vaccination. These chatbot help to reduce vaccine hesitancy and increase vaccination. For example, VWise is a free text-based AI chatbot that guide and inform the public about COVID-19 vaccination in the Eastern Mediterranean Region(EMR) [36].
More than 80% of chatbots had multiple application. The information dissemination and education with the Combating misinformation and fake news (7 cases), The Information dissemination and education with the Self-assessment and screening (6 cases), The Information dissemination and education with the connect to health centers (6 cases), and the self-assessment and screening with the connect to health centers (5 cases) had the highest frequency in the combination of categorized applications.
As shown in Fig. 2, reviewed AI chatbots were different in terms of design features including platforms, AI techniques and user-chatbot interaction design, which are discussed below.
The AI chatbots were deployed on different platforms. Nine chatbots were mobile applications and can be installed on smartphones (Android or IOS). Six chatbots were web-based and can be available through various devices such as computers, tablets, and mobile. Three chatbots were deployed on social media including Facebook (2 cases) and Telegram (1 case). One of the chatbots, in addition to web-based, was also deployed on Facebook[27].
According to our objectives, all of the reviewed chatbots were artificial intelligence based. MIRA and Vac Chat, Fact Check chatbots were hybrid type that used Rule-based and artificial intelligence methods[26, 29]. All of AI chatbots were used Natural Language Understanding (NLU) methods to understand user input and provides the appropriate response to the user in a conversational manner. For instance, WashKaro and Chloe were used LSTM and BERT methods to understand free text input[32, 37]. One of chatbots was used Artificial Intelligence Markup Language (AIML), which is an Extensible Markup Language (XML) based method, to understand user input [7]. Some of chatbots (6 cases) were used machine learning methods such as decision tree and deep learning to identify patterns in user input, make decisions, and learn from past conversations. More than 50% of AI chatbots (9 cases) were used NLU platforms including Google Dialogflow (3 cases), Rasa framework (3 cases) and IBM Watson (3 cases). NLU platforms are prebuilt NLU models that used to development AI chatbots[38].
AI Chatbots had different user-chatbot interaction design. In terms of input format, 13 chatbots were text-based, 1 was voice-based, and 3 cases had both text and voice options. 13 of 16 text-based chatbots were free text input and 3 cases used predefined and controlled text input. In terms of output format, 8 chatbots were able to offer information in multimedia format, including text, audio, image, and video. 5 cases were text-based and 4 cases were text and audio based. In terms of dialog style, 9 chatbots were designed to be user-initiated. In these chatbots user initiate conversation and chatbot response to user’s asks such as questions related to COVID-19. The dialog style of most preventive chatbots (9 cases) were designed in this way. In 5 cases, chatbots were initiator and the user must be answers to the chatbot's questions about things like their demographic data, medical history and COVID-19 symptoms. For example, Ada Health is a mobile-based diagnostic chatbot that asks a series of questions to users screening[33].
In this scoping review, we investigated 17 AI chatbots related to the COVID-19 pandemic. Most of the chatbots used English as the language of conversation with users, while other languages such as German, Chinese, and French were less common. This is consistent with the fact that most chatbots developed in the United States, followed by India and Canada. Chatbots created in India were often multilingual, this is due to the language diversity in this country[7, 24, 25]. Most of the chatbots are designed for the general population and less targeted to specific groups. The generality of most chatbots can be due to the fact that the COVID-19 disease is not related to a specific group and has involved the general population[1]. However, based on the AI chatbots features, this novel technology can be useful for providing better health services to high-risk groups such as elderly people.
Our review found that the eight most common use cases of AI chatbots during the COVID-19 pandemic. Enhancing public awareness about COVID-19 have a significant effect on reducing and preventing the spread of the disease [39]. More than 75% of chatbots have the task of disseminating information and educating users about COVID-19. These chatbots answer users' questions 24/7 and provide users with the necessary information. During the COVID-19 outbreak, misinformation and rumors spread rapidly due to the lack of easily accessible and reliable information from approved sources[37, 40]. Some AI chatbots were developed to fighting this problem by providing reliable information and identifying rumors and fake news[28, 35]. Dissemination of up-to-date, reliable, correct and generally high-quality can be effective in the usefulness of chatbots. Information and findings related to COVID-19 are changing rapidly, so the information provided by chatbots must be updated and validated continuously.
Based on the importance of early identification of people with COVID-19[41], some chatbots were developed to assessment and triage of individuals. These chatbots were often created by governments and organizations such as WHO and CDC to screen people in the community. They usually check the user's risk of contracting COVID-19 based on common diagnostic guidelines and using machine learning techniques. Using these chatbots have many benefits, including reducing unnecessary encounters to healthcare centers, social distancing, efficient use of health resources, preventing virus transmission and early identification of patients. The diagnostic accuracy of chatbot in COVID-19 diagnosis is one of the main challenges of these chatbots. Munsch and his colleagues have shown that chatbots have different sensitivity and specificity in relation to the diagnosis of COVID-19[42]. Therefore, the diagnostic performance of chatbots must be evaluated.
Providing care services to patients who are in home quarantine and continuous monitoring of physical and mental condition of theirs is important in better management of COVID-19[43]. Some chatbots were designed to support COVID-19 patients. These chatbots provide tele-monitoring/care services and allow users to online communicate with healthcare providers and medical centers[33]. This can improve the quality-of-care services, increase patient safety and reduce home quarantine violations.
The COVID-19 vaccines have an effective role in reducing mortality rate and the spread of the virus. However, rumors about the side effects of the COVID-19 vaccines and vaccine hesitancy are main challenges[44]. According to the WHO, vaccine hesitancy is considered one of the 10 major threats of public health[26]. Some chatbots develop to provide services about COVID-19 vaccine such as answering user's questions. These chatbots can be effective in reducing vaccine hesitancy and encouraging people to get vaccinated.
According to the application of AI chatbot in response to COVID-19, AI chatbots can facilitate access to health information and improve quality of healthcare service delivery during pandemic. Equitable access of all people to this novel technology and attaining from its benefits depends on various issues such as access to internet and digital technologies, e-health literacy, age and income status. In addition, investigation and validation of chatbots from effectiveness, efficiency, usefulness and user satisfaction are important determinants that require more studies in these areas.
AI chatbots were also examined from technical aspects. AI chatbots were deployed on various platforms including mobile, web and social media. Mobile-based chatbots were the most frequent. This can be due to the popularity of smart phones and easy access to this technology. 88% of mobile-based chatbots had diagnostic or preventive role and were utilized in self-assessment and dissemination of information. NLU component is essential part of every AI chatbot[38]. Therefore, all of chatbot use NLU methods to understand natural language input and act on the user’s request. As developing an NLU from scratch is very difficult because it requires NLP expertise, chatbot developers usually use NLU platforms[38]. Recently, many NLU platforms were provided to serve as an off-the-shelf NLU component for chatbots. Google DialogFlow, Rasa, IBM Watson, and Microsoft LUIS platforms are the four most common platforms for creating chatbots[38]. In 52% of the reviewed AI chatbots, Rasa framework, IBM Watson and Google DialogFlow platforms were used in. Regardless of the advantages of using the NLU platforms, each of them has features and limitations that should be considered when developing a chatbot[38].
User-chatbot interaction design is one of the important factors in chatbot efficiency[45]. So, we investigated conversational style of chatbots. Input type in the most chatbots was text. In more than 80% of these chatbots, the user can use free text input and have natural language conversation. AI chatbots use various NLP or NLU techniques to process and understand user requests. Therefore, the user can enter his request in the chatbot in natural language and in the form of free text. However, the results of a study have shown that the most of chatbots were not free text and the user could only choose from predetermined and controlled options[46]. This may be due to the most of the reviewed chatbots were rule-based. Input type in Some chatbots were based on voice in addition to text. These chatbots used different voice recognition techniques to understand the input information. For example, 3 chatbots used voice recognition technology from IBM [18, 19] and Amazon [24]. The possibility of user interaction in the form of free text or voice can be improve usability of chatbot and increase ease of use, especially for the older adults and disabled people. However, misunderstanding of user's request and language limitation are major issues in these chatbots. Output format of the most chatbots were text, but other formats such as audio, medical catalogs and educational videos were also used. Using different and appropriate output formats can be help to better and effective user-chatbot conversation, especially in educational and informational chatbots. CoronaGO is an example of these chatbots that provided the necessary training related to self-care against COVID-19 to users in multimedia formats[25]. Chatbot's ability to understand user emotions is important especially in mental health chatbots. With the advancements that are happening in artificial intelligence, it seems that AI chatbots will be able to improve the social and emotional aspects of the conversation.
We reviewed AI chatbots related to COVID-19 and identified 8 common application categories. Most of the developed chatbots have preventive role and commonly utilized for information dissemination, education and self-assessment. Increasing people's awareness, optimal use of health resources, reducing unnecessary encounters, early detection of disease and protection of healthcare providers from exposure to the coronavirus are some advantages of using AI chatbots during COVID-19 outbreak. More than half of the AI chatbots were designed based on widely used NLU platforms. Using NLU platforms can be a suitable solution in the development of AI chatbots in healthcare domain. Because this approach facilitates chatbot development and allows to developers be less involved with technical issues. With advancement in AI domain and NLU platforms, it seems that AI chatbots will mark a bright future in healthcare domain specially in public health, chronic diseases management and mental health.
AI- Artificial intelligence
NLP- natural language processing
WHO- World Health Organization
PRISMA- Preferred Reporting Items for Systematic Reviews and Meta-Analyses
EMR- Eastern Mediterranean Region
NLU- Natural Language Understanding
Ethics approval and consent to participate: IR.ARUMS.MEDICINE.REC.1401.095
Consent for publication: Not applicable
Availability of data and materials: All data generated or analyzed during this study available from the corresponding author on reasonable request.
Competing Interests: The authors declare that they have no competing interests.
Funding: No funding was received to assist with the preparation of this manuscript.
Authors' contributions:
Acknowledgements: Not applicable