The slow rise of artificial intelligence technologies in education seems to be playing out in much the same way that it has in other industries.
A.I. is starting to demonstrate usefulness for specific applications, like helping teachers grade assignments and students sharpen their writing skills. But some technologies are still years away from reality, let alone widespread adoption. And perhaps proceeding cautiously is the wise approach, so that the human touch doesn’t get lost in the adoption of new education technologies.
Those were some of the takeaways from a panel discussion Thursday held at Harvard Business School as part of Boston-based LearnLaunch’s annual edtech conference. Another point of emphasis: teachers shouldn’t be worried about getting supplanted by machines (at least not anytime soon).
“The idea is not to replace humans, it’s to partner with them,” said David Galvin, who leads the investment strategy for IBM Watson’s $100 million “ecosystem” venture fund. The goal is “to enhance what they’re doing, and allow them to do things faster.”
WriteLab, run by Galvin’s fellow panelist, Matthew Ramirez, is an example of this. The Berkeley, CA-based startup’s software uses natural language processing to analyze essays and provide instant feedback to help students polish their writing. The company’s technology also incorporates machine learning and data analytics to try to understand users’ writing style and make more personalized recommendations over time.
“We set the stage for people to make choices about their writing, and then we learn from those choices,” Ramirez said. “There are many ways to revise a sentence; there are many ways to revise a paragraph. And two students may do it differently”—and both ways could be right.
One of the challenges for Ramirez’s company, like any firm working with large datasets and A.I. technologies, is that there’s still a lot of “unstructured, raw data” out there. “We’re in this really interesting gap between getting a hold of all this unstructured data” and making sense of it, he said. “What are we going to use, what sorts of heuristics are we going to use, to set the stage for the data that we collect?”
That’s where IBM leaders think the company’s Watson cognitive computing technology could play an important role. While Watson’s applications for healthcare and enterprise businesses have grabbed many of the headlines, the company is trying to apply the technology to education as well.
IBM is extending Watson APIs—blocks of software code that enable quick deployment of the A.I. technology—to companies and “allowing them to solve the relevant problems in edtech,” Galvin said. “We want to put Watson in the hands of every teacher. That’s the ultimate goal,” he said.
A.I. isn’t the only technology on the 2016 hot lists with an emerging place in education. In the future, Galvin predicts virtual and augmented reality will be useful learning aids. For example, he has come across startups trying to enhance the experience of walking through an art museum by allowing visitors to use virtual reality headsets to “walk into paintings, walk into the era of the painting,” and perhaps have “conversations with the artist.” This could “allow for much greater absorption of content,” Galvin said.
Ramirez said he wants to see software that would analyze the text of a story written by a student and then automatically create images that illustrate it. “Those are new kinds of experiences in education,” he said.
The discussion was a reminder that technology isn’t perfect—glitches are inevitable—and it doesn’t always have the answers. No matter how far A.I. advances, human connections will likely always be crucial to developing students’ minds, as well as their personalities, attitudes, and the way they treat others.
“Everything is happening so quickly,” said audience member Carol Barash, founder and CEO of Story2. “What are we doing, with all this additional work we can do, to make sure that we’re also producing better human beings?”
In response, Ramirez said WriteLab tries to “create an environment where [users] can relax and practice thoughtfulness in their writing.” Ideally, WriteLab helps students practice habits that they can “apply to other parts of their lives,” he said.
Galvin believes students interacting with robots will eventually be commonplace. IBM Watson, for example, is a partner of CogniToys, which makes Internet-connected toy dinosaurs that can talk with children.
Nevertheless, “in the transition to a more intelligent, cognitive world, it’s the responsibility of the human—the teacher—to determine when and when not to be engaged with technology,” Galvin said.
