Take the Blue Pill and Stay in the Matrix

The rise of AI in the classroom


Photo by Markus Spiske / Unsplash

Just as many innovations start, a professor at Georgia Tech, Ashok Goel had a need. As his numbers of students increased, so did their questions. Their questions were same as always, there was just more of them. As message boards, emails, and texts filled up there wasn’t time in the professors day for items he truly wanted to focus on. He needed help. The solution was hiring Jill Watson, a teachers assistant. Jill did a marvelous job answering questions. She posed interesting questions on message boards to engage students in thought provoking discussions. The students responded well.

At the end of the semester the students were told the truth. Jill Watson was not a person. She was a Knowledge Based Artificial Intelligence (KBAI), a smart chat bot. Many students had absolutely no idea they were speaking with a computer. What computer says ‘yep’ in response? The interesting thing here is Watson never had human help yet achieved a 97% accuracy rate. She is supported by IBM’s Watson analytics system (however, IBM wasn’t consulted in developing Jill’s programming). That is the same system that co-hosted an episode of Jeopardy.

By the 3rd year Jill Watson morphed into two different identities. She became Stacy Sisko and Ian Braun. Students knew that Watson wasn’t real but were not told about Stacy and Ian. Students were advised they had several teacher’s assistants. Some were human. Others were not.

Ian stuck to basic questions and answers just like the original Jill Watson did. Stacy focused on giving important information to her 400 enrolled students. Weekly updates, deadlines, and other items were posted. Stacy also summarized the week in her own words. These summaries ranged from what was happening over the week and any interesting conversations. She would also guide students to read encouraging or insightful information in her summaries.

While students were able to identify the AI at a higher rate in the third year, some couldn’t identify which assistants were human.

Al is already moving forward to assist teachers in other ways. Colleges are already looking for AI to grade papers, take attendance, encourage students, all in customizable ways. Some suggest that eventually the AI will become the professor, and the professor will offer support to the AI.

It will be an interesting journey for sure and I am curious how far the rabbit hole goes.

After all, 2.5 million people are friends with Ai Furuse. She is also an AI chat bot.

References
http://www.slate.com/blogs/future_tense/2016/05/10/a_teaching_assistant_at_georgia_tech_was_actually_an_artificial_intelligence.html
https://www.news.gatech.edu/2017/01/09/jill-watson-round-three
https://www.digitaltrends.com/cool-tech/watson-georgia-tech-ta/
https://capturehighered.com/enrollment-services/can-artificial-intelligence-automate-college-admissions/
https://www.edsurge.com/news/2018-04-11-can-ai-help-students-and-colleges-determine-the-best-fit
https://www.thetechedvocate.org/7-roles-for-artificial-intelligence-in-education/
http://journals.sagepub.com/doi/full/10.1177/2332858417749220
https://edtechmagazine.com/higher/article/2017/05/smart-assistants-will-stay-busy-college-campuses