BY PRADEEP KHANNA
The other day, I went to meet someone in downtown Sydney, Australia. On my way, back on the local train, I looked at my mobile to check my emails and found a message asking me whether I would like to meet the person I had just connected with on my LinkedIn network. So, was this some form of artificial intelligence (AI) at play?
Yes! We now live in a brave new world where AI is the next frontier. We keep hearing about bots, chatbots, teacherbots, digital assistants, machine learning, deep learning and many more such words and often wonder what do they mean.
Just like virtual reality (VR), AI has been around for quite some time. In fact, I remember taking AI as a subject when doing my second master’s degree in computer science at University of Technology in Sydney 17 years ago. So, why so much fuss about AI now? AI will reshape how we live and work, but will AI also reshape the way we learn?
ABOUT BOTS, CHATBOTS, TEACHERBOTS AND A.I.
A bot is software that is designed to automate repetitive tasks. Bots have been around for quite some time. An example is use of bots for searching and cataloguing Web pages for search engines. Another example is shopping bots which pull out prices of an item from different vendors from the Internet. Some recent examples are bots making dinner reservations, adding an appointment to the calendar, or fetching and displaying information.
Chatbots are bots that conduct a conversation mirroring potentially a real-life conversation. Chatbots can either be simple rule-based or more sophisticated AI-based. AI-based chatbots get smarter as more interactions take place. The popularity of messaging apps has been lifting the demand for chatbots. Another way to look at the rise of chatbots is the user migration from Web to apps and now from apps to chatbots.
What are teacherbots? Just like a bot or a chatbot, a teacherbot can be a simple rule-based or smart AI-based. Simple rule-based teacherbots can automate simple teaching tasks whereas a smart AI-based teacherbot can become a teaching assistant (TA). A yet smarter AI-based teacherbot can also be personal tutor.
TEACHERBOTS AT WORK
There are two well known instances of teacherbot pilot projects at the University of Edinburgh in the U.K. and Georgia Tech in the U.S.A. The University of Edinburgh teacherbot project was led by its School of Education in collaboration with the School of Informatics and the Edinburgh College of Art. It was launched in April 2015 by Siân Bayne, professor of Digital Education. “Botty,” as this teacherbot was affectionately called, was created to engage on Twitter with students of Edinburgh’s e-learning and Digital Cultures MOOC on Coursera. This MOOC has had 70,000 enrollments across three course runs. The teacherbot’s primary role was to act like a TA. It could answer simple questions on deadlines, course content, etc. It was also able to answer some complex questions as well, based on AI that had been developed on stored tweets with Twitter hashtag #edcmooc. In “Botty’s” case, the students were aware that a teacherbot and not a human being was answering their questions.
Georgia Tech’s teacherbot was developed by Ashok Goel, a professor of Artificial Intelligence at Georgia Tech. Typically, the 300 students at Georgia Tech’s online AI course posted around 10,000 messages in online forums during a semester. Many of these questions were repetitive in nature. This was enough of a driver for Ashok to initiate work on the teacherbot. Leveraging IBM Watson’s technology platform and a databank of 40,000 questions and answers from past semesters, Ashok developed the smart AI overlay for the teacherbot, calling it “Jill Watson. The students were not told that TA “Jill Watson” was a teacherbot.
“Jill Watson” was launched in Jan 2016. As expected, its responses were not very accurate in the beginning, so responses were moderated by the human TAs before posting in the online forums. But by April, it had become sufficiently “intelligent” to answer the questions without human intervention.
The table on the following page compares the two pilot teacherbot projects.
Many factors determine the accuracy level of any AI project, including the AI technology layer at the infrastructure level, the size of the database, and the contextualizing smart AI layer. Looking at the above comparison between the two projects, the Georgia Tech project does have an advantage of using IBM Watson as a technology platform and having a database of 40,000 questions and answers from previous courses. No wonder, it performed better.
AI NEXT IN LEARNING?
The potential of AI to disrupt education and skills training sectors is immense. As Microsoft’s Bill Gates remarked sometime back, we already have online tutoring services where humans provide the services while the platform is online. Smart AI-based teacherbots can replace the humans to provide personalized learning. This has special relevance in lifelong learning scenarios where we will be dipping in and out the learning continuum all through our life.
Automated assessments are a natural application of AI in education and skills training. This application gets further amplified when large number of assessments are being done in an online environment. Use of technologies like Facebook’s facial recognition technology and proctoring are classic examples.
Are we already there in the brave new world where AI is reshaping the way learn? IN these early days, where we are seeing AI-based projects being rolled out in different parts of the world. In the first instance, the focus appears to be on automating routine teaching tasks. This is akin to the Robotic and Process Automation (RPA) implementation onslaught we are seeing in other industries.
“Jill Watson” is estimated to have taken 1500 hours to develop. When many “Jill Watsons” are produced in 15 hours is when we will see real disruption in education and skills training.
Developments in AI in education and skills training will to an extent follow the developments in AI in general. With all major technology innovators investing heavily in AI, it appears certain that our working and learning will get reshaped by AI in future.
—Pradeep Khanna is founder & CEO of Global Mindset (www.globalmindset.com.au) and Technology-enabled Innovations in Learning & Teaching (TILT). He works on enhancing collaborative learning across boundaries and by leveraging technology. Khanna has also been Global Delivery Leader for IBM GBS Australia/New Zealand. He lives in Sydney, Australia, and can be contacted via email at firstname.lastname@example.org