Digital Assistants and Education
In January 2017 a panel of experts ranked artificial intelligence (AI) as the #2 education technology trend to watch in 2017. The trend is important on a number of fronts.
First, students need to be prepared for careers in the AI industry, which will necessitate courses taught on AI, machine learning, natural language processing, etc.
Second, AI offers the potential to improve student support as chatbots and other AI implementation can be used to address student questions involving scheduling, financial aid, and other general student support issues.
Finally, AI is beginning to be used in the classroom as both a means to education and as a support to instructors in their teaching responsibilities. Specifically, using AI in assessment is closer to a real possibility than ever before.
While it is important for educators to think about how to provide students the opportunity to train for careers in AI, the second and third application of AI is germane to a wider swath of the teaching community. Indeed, the use of AI as a teaching assistant to provide student support and to aid in assessment are potentially invaluable tools for the educator.
AI in education is a new frontier in teaching. The opportunities are exciting and innovative educators will benefit from creative use of this growing technology.
AI as Teaching Assistant
Many teachers already use technology in assessment, particularly with online multiple choice assessments or the use of automatic grading of paper-based quizzes using a Scantron machine.
AI also has the potential to go beyond marking static tests using technology to providing substantive assessment to more complex assignments. In particular, AI has the ability to offer feedback to short answer essay questions.
An example of this technology is that developed by [Cognii], a company focusing on the use of natural language processing and education. Cognii’s software is designed to offer ‘virtual tutoring’ that provides students with the opportunity to receive immediate and substantive feedback on short-essay questions.
The technology has matured in the past several years and Cognii recently reached an agreement with Colorado State University’s Psychology department to implement the tool in undergraduate courses.
Watch the video below to see Cognii in action:
Take a moment to reflect on the strengths and weaknesses of using Cognii’s software to grade short essays.
All technologies have unintended consequences. What are some of the unintended consequences that you see with using AI to grade essays.
A use of AI that offers promise to educators is the using AI as a teaching assistant. The goal of such usage is to free teachers from some of the mundane tasks of teaching that can be covered by digital assistants.
It is a challenging task for educators to balance their time between administrative work and the task of teaching. While all educators are keenly aware of this challenge, those who have taught courses with large enrollment feel the pressure acutely.
One of the most obvious pain points in serving a large number of students is offering timely feedback to student questions. Many of these questions are substantive, but, perhaps, more of the questions are mundane.
These sort of mundane questions, like assignment due dates, inquiries about office hours, basic questions about assignments, etc. have the potential to eat up valuable time that could be invested in more substantive ways.
This is exactly the challenge Ashok Goel, Professor of Computer Science at Georgia Institute of Technology, sought to solve in the Spring of 2016.
Faced with teaching a course with 300 students enrolled, Goel and his colleagues, ’trained’ a digital assistant to answer student questions. ‘Jill Watson,’ powered IBM’ s Watson technology, was able to answer 40,000 questions in a database of questions asked in previous courses. The results were so authentic that only a handful of students suspected that ‘Jill’ might not be human.
The lifelike nature of digital assistants will continue to increase with advances in technology thanks to companies like Soul Machines. Founded by professor and entrepreneur Mark Sagar, Soul Machines is developing expressive digital faces for chatbots. Powered by IBM's Watson, the same source as Dr. Goel's teaching assistant, Soul Machines has a live example of their software in use by the Australian government.
For Professor Goel’s summary of the experience using a digital assistant, see the video below:
After watching this video, share your thoughts on what you think would be most significant about using an AI teaching assistant in your context.
Obviously, not everyone one thinks that using an AI teaching assistant is a good idea. What do you think the most detrimental uses of an AI teaching assistant would be?
One interesting response to Goel’s use of Jill Watson is from Roger Schank, former head of the Artificial Intelligence Project at Yale University. Schank argues that the bar is being set too low for AI. What do you think of his critique?