Experts: Artificial intelligence is promising but poses challenges for K-12 education

“Predictive artificial intelligence” like GPS, Siri, and Alexa is a rapidly emerging technology that is slowly making its way into K-12 education — and for good reason, according to the Consortium for School Networking Emerging Technologies Committee.

A new CoSN report, Artificial Intelligence: Could emerging technologies “humanize” teaching & learning?, explores how AI might be used in the classroom as well as the challenges it poses.

The paper defines AI as a device or computer system that performs human intelligence tasks leveraging complex data sets. According to the report’s authors, the most notable examples in today’s schools might include:
· Mixed reality. Students are engaging with an educational platform that combines the physical world and a virtual world, aiming to improve science, technology, engineering, and math skills.
· Augmented reality. “Smart glasses” are being piloted for teachers, which would allow them to assess student progress and compare performance across the entire class in real-time.
· Cognitive tutors. A combination of computer science, cognitive science, and big data is delivering customized instruction to students and new insights to teachers.

Choices and options

Norton Gusky, an educational technology coordinator at NLG Consulting and cochair of the CoSN Emerging Technologies Committee, talked about the report at the CoSN 2018 Annual Conference and offered examples from the Montour School District in Pittsburgh, which uses artificial intelligence tools as educators work directly with researchers from Carnegie Mellon University.

He also discussed IBM’s Watson Education technological tool, which allows teachers to search for lesson plans based on the learning styles, aptitude, and interests of students, as well as the Miami-Dade County Public School District’s use of Carnegie Learning’s MATHia — software that adapts to a child’s mastery level.

“When I first started teaching and I got involved in computers, the first thing I wanted to do was have kids be digital creators,” Gusky said. “This was back in the 1980s, and it was a challenge. I taught a workshop in the 1980s on robotics in the classroom. We were starting to do things, but it was so rudimentary to what we do today.”

Now, AI has the potential to be used in classrooms to create playlists that personalize learning to a student’s needs in a variety of subject areas based on algorithms that leverage “large data sets” — a task that is a heavy lift for time-strapped teachers, he said.

“The amount of time and energy it takes to gather all the data and regroup kids to do differential learning becomes almost next to impossible,” Gusky said. “Now, the technology can do that, but it shouldn’t do it for you; it should do it in a way to give you choices and give you options.”

‘Inscrutable data’

That’s the catch, said Jason Swanson, director of strategic foresight for KnowledgeWorks. Current AI tools are flawed by what he describes as “inscrutable data” that creates “algorithmic bias.”

Put simply, users don’t know how the data is derived; therefore, it can’t be fully trusted, the pair explained.

“The information that we see coming from Google, Amazon, [and] Facebook have been delivered based on algorithms that don’t allow us to see the whole world,” Gusky said.

“We need to have a ‘black box’ solution when these things are doing things like creating playlists and making recommendations for where a student should go,” Swanson added.

Emily Ann Brown covers education technology and STEM education issues for LRP Publications.

Copyright 2018© LRP Publications, Education Daily®