As chatbots are conquering work, there is growing interest in integrating them into workplace training as well. Can chatbots for online learning enhance the educational experience? How does a chatbot facilitate online learning, and what gaps does it fill? Let’s dive into the scientific research to uncover the answers.
tl;dr for Busy People: The Key Benefits of Using Chatbots for Online Learning
- Personalization: Chatbots can provide personalized recommendations for learning resources and mentoring, relieving workers of the burden of choosing suitable learning content. They can also assess learners’ work in a personalized way, delivering qualitative assessments of freely written forms, such as essays and open-question quizzes.
- Encouragement and Learning Support: Chatbots can facilitate positive emotional engagement with learning tasks by using visual and textual elements such as humor, jokes, emojis, animated gifs, and pop culture references. They also have the potential to improve engagement through gamification techniques, for instance, by sending personalized messages that work better than classic reminders, nudging employees to remember their learning objectives.
- Facilitation of Information Transfer: Chatbots can provide learning opportunities that are not restricted by space or time. They can handle multiple questions at once, saving users’ time for other tasks. Chatbots can also automate information transfer, lessening lecturers’ administrative workload and improving the efficiency of training processes. They enhance access to information and promote better comprehension and satisfaction with the learning process.
- Cost-Effectiveness and Automation: Chatbots provide scalable learning operations that are less costly than traditional methods. They offer a safe learning environment, particularly beneficial for novice learners. They can also play a significant role in facilitating the onboarding process for new hires by providing personalized learning content recommendations.
- Versatility and Accessibility: Chatbots use intuitive interfaces of popular chat applications, making them easy to use and widely accessible. They can also help accommodate employees with different learning disabilities.
- Reduction in Administrative Load: By handling routine and repetitive administrative tasks, chatbots allow trainers and administrators to focus more on curriculum development, research, and higher-level tasks, leading to increased efficiency.
- Despite these benefits, the article also mentions potential limitations and caveats, such as the potential for chatbots to distract rather than support, limitations in user-friendly design, and the possibility of incorrect information being given by AI-based chatbots like ChatGPT.
- However, the overall impact on workplace learning seems to be positive, particularly in the context of quick employee upskilling activities.
Learning in the Age of Digital Transformation
Today, digital transformation is a fact — such a transition impacts where we do our business (mostly online) and how. Productivity rates in Industry 4.0, as experts call this phase of technological evolution, are determined by complex machines and software systems rather than human labor. Consequently, human supervisors of those intelligent structures have been tasked with a new set of duties focused on control and optimization. Such a shift in the employee competence landscape required profound re-learning initiatives. How we train our employees had to change as well. Daily collaboration with intelligent systems in the workplace prompts many to introduce new methods to workplace training. One of the pioneering methodologies includes the use of chatbots.
History of Chatbots
Chatbots, a new alternative to traditional online learning, are not as new as we think. The first chatbot was developed in 1966 by Joseph Weizenbaum at MIT. ELIZA was supposed to act as a psychotherapist, and her responses were based on a template through pattern-matching techniques to respond to the user’s query (Brandtzaeg and Følstad, 2017, pp. 377–392)” (Okonkwo and Ade-Ibijola, 2021, p. 2).
The first chatbot to pass Turing’s Test (a test where users cannot distinguish whether they talk to a human or a machine) was A.L.I.C.E. or Artificial Linguistic Internet Computer Entity, developed in 1995 by Richard Wallace. ALICE won the Loebner Prize and became known as the first “human-computer.” Wallace released ALICE as open-source software, changing the landscape of the chatbot, or chatterbot, software market, allowing software developers to use AIML (Artificial Intelligence Markup Language) in their own work.
Since then, the development of so-called virtual assistants burgeoned first with Mitsuku and Jabberwacky (a British nod to a classic Monty Python comedy), then Apple Siri, Amazon Alexa, and Google Assistant that serve us today, all marking significant milestones in AI history, until ChatGPT arrived, heralding a new era with an unprecedented depth of contextual understanding and conversational versatility.
What Is a Chatbot?
Both industry and academic literature coined numerous terms to call it: a pedagogical agent, a conversational companion, a voice or cognitive assistant, or a virtual buddy are just a few that we found. In literature, a chatbot is „a software tool that interacts with users on a certain topic or in a specific domain in a natural, conversational way using text and voice” (Smutny and Schreiberova, 2020, p. 1). Thanks to its scripted or AI-sourced ability to respond to the users’ input, a chatbot is capable of providing assistance across various domains.
Do we want chatbots to pretend to be humans? Seemingly not. According to research, “Rather than creating a human-like smart machine application, it is about creating effective digital assistants who are able to provide information, answer questions, discuss a specific topic, or perform a task.” (Smutny and Schreiberova, 2020, p. 1). Consequently, education and learning has become one of the chief use cases for chatbot assistance
Chatbots as Pedagogical Agents
According to Clarizia et al., “Chatbots are being considered as a useful technology to facilitate learning within the educational context (Clarizia et al., 2018, pp. 291–302)”. Okonkwo and Ade-Ibijola claim that “Chatbots are increasingly being used to improve student interaction in this current world of technology where communication and many other activities rely heavily on online platforms” (Okonkwo and Ade-Ibijola, 2021, p. 1). Suppose we understand chatbots as automated response programs that interact with users through a conversational interface, providing them with information based on written or voice prompts (questions or queries). In that case, chatbots can act as educators and revolutionize education.
In their article, Smutny and Schreiberova mention that “Chatbots have had a long history of use as pedagogical agents in educational settings. Since the early 1970s, pedagogical agents within digital learning environments known as Intelligent Tutoring Systems have been developed (Laurillard, 2013)” (Smutny and Schreiberova, 2020, p. 2) and Colace et al. state that “Chatbot technology can be considered an important innovation for e-learning: in fact, they are turned out to be the most innovative solution in filling the gap between technology and education” (Colace et al., 2018, p. 528).
Workplace Learning in the Digital Age
Digitalization of the learning experience revolutionized when, where, and how we absorb knowledge. In the Digital Age, our access to mobile devices and internet connectivity almost 24/7 enables us to access all kinds of knowledge and tools at any point, anywhere. But how is digital education different than traditional one? Specialists from the Institute for Academic Development at the University of Edinburgh define digital education as „the innovative use of digital tools and technologies during teaching and learning.” Digital education is often „referred to as Technology Enhanced Learning (TEL) or e-Learning. Exploring the use of digital technologies gives educators the opportunity to design engaging learning opportunities […] and these can take the form of blended or fully online courses and programs”.
In the workplace, digital education or eLearning has been adopted quite early. It offered scalability, flexibility, and cost-efficiency that traditional training couldn’t. Online learning tools allow for the creation and spread of learning materials at a rate never seen before. Digital documents can easily be updated and remodeled. Learners can complete the training at their own pace, in their own time. What other improvements could eLearning need?
E-Learning: A Step Forward but in the Right Direction?
The range of digital learning tools available to teachers, trainers and learners constantly expands. Most of them are available on mobile devices, where our daily knowledge absorption occurs. This is where our other most important daily activity takes place: conversations. As Feng claims, “Informal learning, i.e., learning “from daily life activities related to work, family, or leisure” (Colardyn and Bjornavold 2004, p. 71), has proven to be an integral part of workplace learning (Billett 2004; Rogers 2014; Tynjälä 2008), facilitating the development of soft and hard skills through human-to-human interactions and discussions.” (Feng 2022, p. 2).
In the age of remote and hybrid working models and nomadic lifestyles of digital entrepreneurs and content creators, it’s hard to retain connections similar to those at a physical workplace. The serendipity of lunchbreak, lift, or passing-by chats is difficult to emulate in the somehow desolate environment of our home office spaces. Traditional e-Learning is still appreciated by many. But with the introduction of intelligent machines and complex systems, which require constant upskilling from their operators, traditional eLearning materials stopped serving their purpose. Much of what we do now at work, we have never done before. Actionable e-Learning materials take too long to produce, review, and release. They are simply too slow to address immediate needs for case-based troubleshooting.
The Rise of Conversational Microlearning
Conversational learning is an innovative approach to learning that taps into our most natural way of gaining new knowledge – asking questions, being asked, and coming to conclusions together. Conversational agents engage in a short and to-the-point chat-based conversation with the learner aimed at teaching or refreshing their knowledge. Lessons usually don’t take longer than 2 minutes and are organized into a pre-scripted or AI-enhanced question-answer exchange, where first the chatbot assesses the student’s existing knowledge and then engages them in a quick exercise. Learn more about conversational microlearning. According to Kuhail, “The adoption of educational chatbots is on the rise due to their ability to provide a cost-effective method to engage students and provide a personalized learning experience (Benotti et al., 2018)” (Kuhail et al., 2023, p. 974).
Chatbots have the ability to emulate human-like conversations that resemble casual chats with coworkers. Using humor, emojis, and references to popular culture, chatbots step in where in-person social connections have been severed. Chatbots also provide services that traditional eLearning lacks, such as individual support and feedback. In increasingly digitalized workplaces, chatbots, themselves software-based tools, assist with introducing and exposing employees to new types of human-computer interactions.
How Chatbots Facilitate Online Learning
To answer that question, we have reviewed the latest studies in the area of computer-assisted learning in the context of higher education as well as workplace training. Recent studies show that the use of chatbots as pedagogical agents is rising. Many authors claim that chatbots have proven helpful to students at any level, from kindergarten to college and lifelong learning. For instance, chatbots have helped students while learning languages, reminded them about assignments, and helped them orient themselves during the first few weeks of the new semester. Available studies tackling the issue of introducing chatbots into workplace training also suggest that chatbots, or personal assistants, provide aid, especially during onboarding processes. They were useful not only by submitting the necessary information but also by serving as personal assistants: for instance, by sending regular reminders to complete the training tasks.
The Main Benefits of Chatbots for Online Learning
Overall, the literature is unanimous in that introducing chatbots into the blended learning experience has been largely beneficial. After closely analyzing all the fields where chatbots have proven useful, we would like to focus on four aspects of chatbot-supported training: personalization, encouragement and support, availability and facilitation of information transfer, and cost-efficiency and automation.
Why these four? To be sure, the literature lists many categorizations of chatbot affordances: time efficiency, ease of use, effective tutoring, etc. These four, though, seem to have the most significant impact on improving the individual learning experience (chatbot as a personal tutor) and the efficiency of the business training model in general (cost-efficiency and scalability). Other categories, such as ease of use, are mentioned as elements of the main four used in this article.
Many studies show that a successful learning experience (regarding emotional and educational workload) largely depends on individual support. Providing every learner with that kind of personalization and individual approach remains sadly unattainable, whether in education or the business environment. Conversational agents can fill this void. According to Feng (2022), chatbots can act as personalized recommenders: “A chatbot can access data sources to generate a personalized recommendation, e.g. learning resources or human mentors (Chandar et al. 2017; Liao et al. 2016). Supplying fitting recommendations takes the burden of choice from the worker, to choose suitable learning content, thus lowering the bar towards upskilling.” (2022, p. 1). Furthermore, as Feng suggests, thanks to human-like features, such as empathy, “a chatbot can reach users on a more personal level than a standard learning program. It feels more natural for a chatbot to organize informal learning opportunities like ‘coffee meetings.’ (Feng 2022, p. 2).
Personalized assessment is another sphere where chatbots can prove useful. Traditional online learning solutions rarely offer individual feedback – mainly due to the quantitative assessment methods and automated scoring systems. Chatbots are capable of delivering a qualitative assessment of freely written forms, such as essays and open-question quizzes. Sabiq and Fahmi conclude, that “Assisting teaching with the technical advances of artificial intelligence in the use of mobile devices will provide students with practical self-guided learning. With the popularity of WhatsApp being a learning tool nowadays, its features can be inserted with chatbots that can offer several advantages to be a teacher assistant, particularly for mediating assessments” (Sabiq and Fahmi, 2020, p. 18).
2. Encouragement and Learning Support
Chatbots meet students where they are, mainly in the ubiquitous chat applications where they can provide human-like emotional support and facilitate positive emotional engagement with the learning task. According to Smutny and Schreiberiva, “chatbots demonstrate the ability to create easygoing interactions with users so that they can be leveraged to support engagement, as well as setting out goals, strategies, and outcomes of learning and training (Cinglevue, 2017)” (Smutny and Schreiberova, 2020, p. 2). They utilize all the visual and textual elements to engage with the student naturally, using humor, jokes, emojis, animated gifs, and references to popular culture. The authors conclude, “These dynamic elements go a long way towards imbuing chatbots with personality. They make the conversation more enjoyable, more immersive, and more visually engaging. Emojis are another way to add personality in an entertaining and evocative way” (Smutny and Schreiberova, 2020, p. 10). According to Huang, “Chatbots appeared to encourage students’ social presence by affective, open, and coherent communication” (Huang et al., 2022, p. 237).
Another way to improve engagement among learners is through gamification techniques. Many LMS systems already use push notifications to remind learners about their objectives (such as finishing the course). The research found that short personalized messages from chatbots work better than classic reminders. Usually, completing a course has low priority on employees’ to-do lists. This digital nudging helps them remember their learning objectives and creates positive links with educational efforts. Less than a nag, reminders sent by chatbots seem to encourage rather than force employees to follow up on the lesson.
3. Facilitation of Information Transfer
Undoubtedly, digital learning removes geospatial boundaries to learning, making it available anytime and anywhere. As Nenkow et al. point out, educational services have been dissociated from geographical and temporal constraints, and “[the use of] LMS today enables us to break the space-time linking between a teacher and a student” (Nenkov et al., p. 192). Unlike teachers and trainers, chatbots are available 24/7, holidays and working hours notwithstanding. According to Okonkwo and Ade-Obijola, chatbots are able to provide immediate assistance by delivering rapid replies to the students’ questions and „have the ability to handle multiple questions at the same time, saving the user time to do other tasks” (Okonkwo and Ade-Ibijola, 2021, p. 6).
The same authors suggest that “In the education domain, chatbots are used not only to develop students’ interaction skills but also to assist teaching faculty by bringing automation (Dsouza et al., 2019). These systems can not only improve student engagement and support, but they can also greatly lessen lecturers’ administrative workload, allowing them to focus on curriculum development and research” (Okonkwo and Ade-Ibijola, 2021, p. 1). In this case, for instance, the teachers wouldn’t have to supervise the introduction classes to the subject, as all the information would be delivered via chatbots. Also, chatbots could generate assessments of short essays or quizzes (btw, eggheads allows you to generate chat-based tutorials and quizzes based on your material).
In a globalized business environment, automation of information transfer plays a massive role in improving the efficiency of training and production processes. Learning materials can be prepared in advance, stored, and distributed automatically. Chatbots send learners push notifications as new resources are uploaded and remind them if their tasks remain incomplete. Thus, chatbots improve accessibility to information, and by using a conversational method of knowledge transfer, which is much easier to comprehend and digest, they enhance efficacy and satisfaction with the learning process. Most importantly, chatbots are easy to use as they use intuitive interfaces of popular chat applications.
4. Cost-Effectiveness and Automation
One of the key benefits of using chatbots for training is the scalability of learning operations. As Feng writes, “Despite human coaching providing reliable results, it does not scale well and is a costly endeavor for schools. […]”. She summarizes particular use cases for chatbots in industry learning: “Hernández et al. (2016) propose a virtual, embodied conversational agent (VECA) with anthropomorphic visuals used in a simulation training for workers in an electric utility company. Conducting training activities in a real, hands-on training bears life-threatening risks when the trainee is a novice, so the VECA setup provides a safe training environment.
Subramanian et al. (2019), Casillo et al. (2020) and Dominic et al. (2020) propose chatbots as intelligent recommenders for learning content for information workers. By fulfilling this role, the objective is saving time for the learner and provide better fitting content. This facilitates the onboarding phase of the new hires and new contributors (Casillo et al. 2020; Dominic et al. 2020) or enable software engineers to maintain continuous learning (Subramanian et al. 2019). In addition to a faster and better onboarding experience, Dominic et al. (2020) point out that it can increase the retention rate of newcomers to open-source software projects.” (Fend 2022, p. 4).
To sum up, chatbots are invaluable in employee training, especially in its early stages. Not only do they deliver material when and where it’s convenient for the learners, but they also send individual recommendations and allow the learners to set the pace of the course, often saving their time. Most importantly, they create a safe learning environment, especially for the novice learners.
Chatbots as Educational Agents: Pros and Cons
Using chatbots as pedagogical agents also has its downsides. Most importantly, there is limited evidence on the long-lasting effects of using chatbots. Thus far, most experiments focused on the short-term use of chatbots for educational purposes. Kuhail et al. note that “Fryer et al. (2017) found that students’ interest in communicating with the chatbot significantly dropped in a longitudinal study. The decline happened between the first and the second tasks suggesting a novelty effect while interacting with the chatbot. Such a decline did not happen when students were interacting with a human partner” (Kuhail et al., 2023, p. 1005). The same study suggests that over time, chatbots presented themselves rather as a distraction than support because their interface was not as user-friendly as expected. Many chatbots analyzed in the studies also lacked usability and human-centered design. What also needs mentioning is that chatbot tutors are as effective as their designers. Poorly designed conversations will leave the learners dissatisfied and possibly prejudiced toward trying new educational technologies.
ChatGPT Propels Chatbots Forward
ChatGPT, the latest breakthrough in the chatbot field, distinguishes itself from earlier models such as ELIZA and ALICE. While these earlier generations operate based on algorithms selecting pre-scripted responses, ChatGPT leverages advanced machine learning technology rooted in Natural Language Processing (NLP). It’s recognized as a Large Language Model (LLM), a subset of AI that is trained on an extensive array of internet text. However, it’s important to note that ChatGPT doesn’t know specifics about which documents were in its training set and it doesn’t have access to any information or updates beyond its last training cut-off in September 2021.
This robust training enables ChatGPT to craft human-like responses while maintaining the context of a conversation, breaking away from the rigid question-answer structure commonly seen in older chatbots. However, ChatGPT doesn’t have the ability to access or search the internet in real-time (yet). To do so, you would need a subscription with OpenAI for ChatGPT Plus to use the Bing Plugin or use Bing Chat.
Despite its potential as an interactive, dynamic learning tool, researchers caution against over-reliance on AI-powered chatbots like ChatGPT. Tlili et al. (2023, p. 8) underscore the risk of learners exploiting these chatbots for quick answers, which could stunt the development of their own innovative capacities and critical thinking skills. “Conversely, a few of the participants held an opposing view that the abuse of ChatGPT by learners can also diminish their innovative capacities and critical thinking. For instance, when learners are not motivated, the probability of seeking an easy-to-get solution is high as can be deducted from a statement from one participant.” (Tlili et al., 2023, p. 8). This is a vital reminder that AI is a tool to aid learning, not to replace the critical and creative thought processes at the core of education.
Additionally, a phenomenon known as ‘hallucinations’ can occur with LLMs like ChatGPT. This term refers to instances where the AI fabricates or fantasizes parts of its responses, resulting in statements that are plausible-sounding but completely false. It is essential for users to be aware of this potential and remain critically engaged, verifying information generated by such models independently. Remember that, despite its impressive conversational abilities, ChatGPT does not ‘know’ or understand information as humans do – it doesn’t form beliefs, have intentions, or experience consciousness. As such, it cannot definitively discern or guarantee the truth of the information it generates.
Limitations and Strengths of Chatbots in Employee Upskilling
While substantiated, the critique above must be put in the proper context. As with every tool, chatbots have certain limitations, and their applicability depends on the use case. There is still no scientific evidence on the implication of the long-term use of chatbots on educational processes and outcomes. While that data is being collected and analyzed, the application of chatbots in quick employee upskilling activities proves highly effective. According to Casillo et al., “One of the main advantages of the proposed system is the simplification of learning and the [time] reduction of this process” (Casilio et al., 2020 p. 5) especially among the new employees, who receive a knowledge-in-a-nutshell orientation training. Moreover, chatbots can provide accommodation for employees with different learning disabilities. Insights gained from the use of conversational agents in such domains as customer relations or cyber security awareness training show, that short lessons delivered via chatbots are user-friendly, efficient, and don’t overburden the learners. Using chatGPT can prove extremely useful when conducted with proper query techniques. Companies could provide training on the constructive use of such resources.
Summary: The Benefits of Chatbots for Online Learning According to Science
Overall, chatbots can be extremely helpful in designing new models of digital learning. They can assist kindergarten teachers, college students, and high-profile employees alike. They are flexible, user-oriented, and scalable. In comparison to traditional models of eLearning, using chatbots improves the automation of training processes, creates a more personalized learning experience, and facilitates systematic repetition (digital nudging), leading to better knowledge retention. Research suggests that new employees significantly benefit from using chatbots during onboarding. Chatbots provide individual attention, visibility, and personalized feedback new employees need at the beginning of their paths. Automation of information transfer is cost and time efficient and makes the learning efforts more scalable compared to traditional online learning models.
eggheads – Chatbots for Online Learning
At eggheads, we realized that meeting learners where they are, that is, on communication apps, providing small nuggets of knowledge at a time, and giving them control over time and space where they learn, brings considerable benefits in knowledge retention and employee learning satisfaction. Educating educators is crucial for us: we assist our clients with conversational design and supply tutorials on maximizing the effectiveness of chatbot lessons. Moreover, designing a new tool is an iterative process in which ideas are tested, applied, and modified to answer the user’s needs. Using human-centered design principles to develop online tools (such as chatbots in education) might substantially improve their usability scores.
Casillo, M., Colace, F., Fabbri, L., Lombardi, M., Romano, A., & Santaniello, D. (2020, December). Chatbot in Industry 4.0: An Approach for Training New Employees. 2020 IEEE International Conference on Teaching, Assessment, and Learning for Engineering (TALE) (pp. 371-376).
Clarizia, F., Colace, F., Lombardi, M., Pascale, F., Santaniello, D. (2018). Chatbot: An Education Support System for Student. In: Castiglione, A., Pop, F., Ficco, M., Palmieri, F. (eds) Cyberspace Safety and Security. CSS 2018. Lecture Notes in Computer Science(), vol 11161. Springer, Cham.
Colace, F., De Santo, M., Lombardi, M., Pascale, F., Pietrosanto, A., & Lemma, S. (2018). Chatbot for e-learning: A case of study. International Journal of Mechanical Engineering and Robotics Research, 7(5), 528-533.
Huang, W., Hew, K. F., & Fryer, L. K. (2022). Chatbots for language learning—Are they really useful? A systematic review of chatbot‐supported language learning. Journal of Computer Assisted Learning, 38(1), 237-257.
Khosrawi-Rad, B., Schlimbach, R., Strohmann, Y., & Robra-Bissantz, S (2022). Design Knowledge for Virtual Learning Companions. Proceedings of the 2022 AIS SIGED International Conference on Information Systems Education and Research. 6.
Mora, A. M., Guillén, A., Barranco, F., Castillo, P. A., & Merelo, J. J. (2021). Studying how to apply chatbots technology in higher education: first results and future strategies. Learning and Collaboration Technologies: Games and Virtual Environments for Learning: 8th International Conference, LCT 2021, Held as Part of the 23rd HCI International Conference, HCII 2021, Virtual Event, July 24–29, 2021, Proceedings, Part II (pp. 185-198).
Nenkov, N., Dimitrov, G., Dyachenko, Y., & Koeva, K. (2016, September). Artificial intelligence technologies for personnel learning management systems. 2016 IEEE 8th International Conference on Intelligent Systems (Is) (pp. 189-195). IEEE.
Sabiq, A. H. A., & Fahmi, M. I. (2020). Mediating Quizzes as Assessment Tool through WhatsApp Auto-response in ELT Online Class. Langkawi: Journal of The Association for Arabic and English, 6(2), pp. 186-201
Tlili, A., Shehata, B., Adarkwah, M. A., Bozkurt, A., Hickey, D. T., Huang, R., & Agyemang, B. (2023). What if the devil is my guardian angel: ChatGPT as a case study of using chatbots in education. Smart Learning Environments, 10(15), pp. 1-24.