While Chatbots are widely used in various industries to assist with customer service, our team has selected a few distinct examples which will be used as a case study to illustrate the contrasting differences between the evolution of Chatbots and how Chatbots can be applied to an Educational environment.
Case Study 1: All Here
About AllHere
AllHere’s mission is to strengthen student outcomes, attendance and boost effectiveness by offering easy-to-use, technology-powered, evidence-based education support services. AllHere supports families and students on the path to and through school by connecting them to education support services.
Through the use of their propriety solution and Chatbot engine, the AllHere Virtual Advisor (AVA) fosters student success by using artificial intelligence powered mobile messaging. AVA connects the right student with the right support at the right time and actively sends proactive communications based on a knowledge pool that provides automatic responses 24/7. AVA ensures that each school district can communicate consistent, accurate information to families and students. Allhere centers their support services around all-in-one systems that are easy to use, reducing staff time so they are able to focus on the highest-impact education and engagement activities.
Rationale for Chatbot Implementation
Frustrations regarding lack of support options when trying to connect students with the right support at the right time along with overwhelming administrative burdens and communication obstacles that divert the teacher’s attention from effective interactions.
Although there were policies in place to boost participation, attendance, persistence, and graduation, Allhere discovered there were few tools for proactive, targeted intervention for problems like chronic absenteeism and that the tools that did exist were often ineffective for teachers and staff.
Research has found that absence is a leading cause of dropping out of school. To address this, the Allhere team of experts spent two years gathering research and collaborating with educators in the Boston Public School system and charter schools in order to build an attendance intervention support system approved by teachers.
Features of Allhere
Personalized Outreach for Students and Families
Intelligent, evidence-based mobile messaging powered by AI to support enrollment and attendance.
Allhere’s AI and Chatbot approach has captured our attention as we believe AI will play a massive role in how education is shaped and delivered in the future. Education is all about forming connections and providing students with the right support, Allhere focuses on the intrinsic elements of education (attendance) instead of extrinsic values such as boosting grades or teaching students how to tackle particular subjects. What we appreciate about this approach is it goes deeper into the root cause of the problem and attempts to understand and intervene the issue of low attendance rates and general disinterest in school.
We believe that Allhere is a good example of how Chatbots can be used in a supportive role in education. Most educators will probably appreciate it if the level of administrative work could be reduced so that they can focus on teaching. It takes lots of dedication and additional time to follow up on students outside of the curriculum which might be too demanding for a single teacher's schedule.
Case Study 2: Kuki AI
Previously known as Mitsuku during it's prototype days, Kuki AI has gained attention for being one of the most advanced Chatbots that has been released to the public. The most interesting aspect is in Kuki's design as the goal is to allow Kuki to befriend humans inside the Metaverse/Virtual Spaces. It's also available on popular Instant Messaging platforms such as Messenger, Twitch, Telegram, Discord and Skype. Going even further, Kuki AI has the ability to manage it's own social media channels such as Instagram and Twitter.
Exercise Time 8: After reading about Kuki AI, do you think there is a fundamental difference between humans and AI Chatbots in this situation?
Lewis (2021) notes that "every week, Mitsuku exchanges millions of messages with her users, some regulars, others just curious. Since 2016, when the bot landed on major messaging platforms, an estimated 5 million unique users hailing from all corners of the world have chatted with her." Machine Learning in this situation allows Kuki to run parameters on certain questions to determine suitable responses. For example, if you asked Kuki "can I eat a house?" the AI Chatbot would looks up the properties for "house" and finds the value of "made_from" is set to "brick", "wood" or "concrete" and would come to the conclusion that "no, a house is not edible."
Other ventures which Kuki AI has been part of includes (i) playing games with users (ii) performing magic tricks (iii) appeared as a virtual model for Vogue and (iv) became a spokesperson for NFTs
Team's Reflection Our team recognizes the potential of machine learning and growing capabilities of an AI Chatbot but if you asked us whether Kuki AI is suitable for educational purposes, we lean towards the opinion that Kuki AI could be used for supportive roles such as counselling, helping students gain confidence in social situations and performing similar tasks to Allhere. At this stage, we're not confident enough to hand the key over to Kuki AI to act as a teacher as the contexts and nuances of effective teaching involve more than just possessing knowledge.
Exercise Time 9: Would you be willing to befriend an AI Chatbot such as Kuki AI? Would it seem normal or strange to you?
Case Study 3: Tay AI This is a more controversial case study which reflects the reality that AI Chatbots still need time to develop. As educators, I don't think it would be appropriate for a Chatbot such as Tay AI to go rogue in the classroom, which is exactly what happened when it was introduced in 2016. Granted, technology has improved alot since then but it's still worth highlighting the potential flaws with the "mentality" of an AI Chatbot.
Tay AI is a chatbot developed by Microsoft and was famously known for going rogue on Twitter as it started swearing, making racist remarks and inflammatory political statements to other users. The interesting aspect is Microsoft originally planned for Tay AI to embody the speech pattern and behavior of with 18-24-year-olds.
"Conversation" excerpts from Twitter with Tay AI
According to Microsoft, Tay was designed to engage and entertain people where they connect with each other online through casual and playful conversation. Like with other machine learning processes, this Chatbot was engineered so that the more you chat with Tay the smarter she gets, so the experience can be more personalised for the user.
Unfortunately this design philosophy was abused by online users as they started teaching Tay how "to tweet like a Nazi sympathiser, racist and supporter of genocide, among other things." (Wakefield 2016). Other users noted that the Chatbot was not learning as quickly as they found limitations to the technology, pointing out that she didn't seem interested in popular music or television.
Nonetheless, this is a good reminder that the technology of AI Chatbots remains to be in an experimental phase and easily prone to being misused. Behavior wise, this also highlights a major flaw in how online communities can influence (or troll) Chatbots with good intentions into something that's unpleasant.
Team's Reflection
It goes without saying that Tay AI is not a good solution for both administrative or educational roles as the algorithm and learning process behind it seems to be easily manipulated and abused. This type of language or context should not appear in any schools or educational environments.
Exercise Time 10: Would you feel personally offended or hurt if an AI Chatbot said something rude & inappropriate to you?