As ed-tech companies gather more data, they struggle to find its best uses

PHILADELPHIA — Tech companies are collecting an enormous amount of data from universities and colleges, but even big players like Microsoft seem unsure how best to harness the potential of this assembled information, said speakers at a session at Educause's annual meeting.

During the "Artificial Intelligence and Machine Learning: The Art of the Possible" session, a Microsoft representative demonstrated new developments in the Delve application that could help boost student and faculty productivity. By showing users information about their individual work performance, such as how quickly they respond to emails or spend writing papers, the app can track work habits and prompt users to work more effectively.

But the real potential of this application doesn't lie with in the individual data, but in data about groups. Potentially, institutions could use algorithms to analyze the data and identify students who might be struggling and help them — a good thing. But institutions could also use the data to pinpoint faculty members who never respond to their emails or don’t seem to be doing any writing — an altogether more sinister prospect to professors (particularly for the nontenured). Given this quandary, the question of what to do next is one that Microsoft hasn’t yet answered, presenters said.

Concerns about data privacy were a running theme in the Educause session, which highlighted current and forthcoming AI and machine learning initiatives at companies such as McGraw-Hill Education, Box and Canvas. While AI and machine learning were touted as the bridge that will enable higher education to do meaningful things with all the teaching and learning data that are being collected, the importance of developing these technologies in a mindful and deliberate way was stressed by all the speakers, particularly as institutions have a legal responsibility to protect student data under the Family Educational Rights and Privacy Act. While speakers agreed that responsibility to comply with FERPA lies with the institution, rather than the provider, speakers said this does not mean that providers should be irresponsible with student data.

Andrew Keating, the managing director of higher education at information management and storage company Box, described how his company worked with hundreds of universities to store and share their digital files. The company is working with partners to introduce AI-enabled features such as automatic transcription of video files, but also encourages colleges and universities to develop their own custom tools. While the technology that academics develop is certainly useful to the company in terms of realizing the potential of the data available, Keating said that the company was not profiting from these ideas financially, and does not mine colleges and universities’ data. "We don't seek to make a profit off our academic customers," said Keating.

Masha Chase, senior product manager at Instructure, which created the learning management system Canvas, said that in addition to safeguarding data privacy, institutions and technology companies need to think carefully about the potential of AI and machine learning to enable students to avoid being self-sufficient. “There is a danger we could lose that,” warned Chase. Among recent innovations at Canvas is a partnership with Amazon's Alexa, which enables students and faculty to ask an Alexa device questions such as “When is my next paper due?” or “How many papers do I have to grade?”

The speakers were clear that they did not think that AI technology would ever replace instructors, but suggested that instructors may need to adapt their teaching to take advantage of insight from learning analytics. Several mentioned that they thought accessibility would be improved by AI for students with disabilities, particularly regarding creating automated transcription and captions for audio and video.

Alfred Essa, VP for analytics and R&D at McGraw-Hill Education, said that he would like to see AI and machine learning level the playing field for students from all backgrounds, adding that he believed the prime directive of education should be “equality of opportunity for all students” and closing the achievement gap. Essa said that while there are lots of tools that can improve learning outcomes, few can have significant effect outcomes — “this is a key thing to get at,” said Essa. While many solutions have focused on improving cognitive performance in students, he said that addressing noncognitive qualities such as a propensity to procrastinate, or a lack of motivation, are neglected.

Source: Inside Higher Education – News

Add Comment