Julian, powered by Google, teaches and learns at Walden University
Steven Tom, Chief Client Officer at Adtalem Global Education, was at a conference several years ago and saw a demonstration from an AI tutor that left him thinking bigger. The tool he saw took students’ questions and answered them according to what he called a script, following a predefined path programmed by a human in the back which, according to the idea, ultimately led the student to the correct answer. It sounded a lot like adaptive learning, a concept that’s been around for decades and has failed to take off, in part because it takes a lot of effort to program and is quite inflexible in its response.
“Someone must have spent hours and hours and days and days creating the questions, scripting how the tutor was going to interact with the student,” says Tom. “And then the AI part was like when you read a news article and gave it a thumbs up, thumbs down — the type of learning based on whether you liked the question or not. In that sense , it really wasn’t dynamic and it really wasn’t scalable.
This experience set Tom on a mission to fix both of these weaknesses.
His first step was to reach out to Google teams working in AI and higher education. Tom’s team at Walden University, an online university recently acquired by Adtalem, worked with Google to create the AI tutor that would be introduced to students and faculty as Julian in spring 2021.
“We wanted to see if we could meet this challenge of creating a really dynamic, really unscripted AI tutor that could essentially ingest content on its own, make sense of it, extract key concepts from it, and then nurture with the studying a real tutorial session where he can generate his own questions and assess the students’ answers completely on his own,” says Tom. “At the time, it was kind of a pipe dream.”
Today, says Tom, that dream has at least partly come true.
Julian is more dynamic than most AI tutors. By design, it’s been deployed in lessons where it’s not easy for a machine to tell if the answer is right or wrong (as opposed to, say, a math lesson). At Walden, Julian’s first courses were in early childhood education and sociology.
DISCOVER: Successful examples of AI in higher education that can inspire our future.
And while Julian doesn’t exactly judge right or wrong – he doesn’t do any grading, for example – he ingests information and learns as he goes. It uses the same course material that students receive and refers them to these sources in response to questions they ask. Then it takes those questions and answers and feeds them back to learn what students are asking and get even smarter for next time, Tom says.
Because it was developed with Google, Julian lives in Google Cloud, which meant little to no infrastructure investment was required from Walden. Tom says the programmers opted to use an application programming interface to drive the tool, keeping an eye on its future use.
Because it’s API-driven, Tom says, Julian can potentially appear in many environments: while it’s currently a chatbot integrated into the university’s learning management system, it could easily exist as an avatar in an augmented or virtual reality environment, or as a voice. on the phone in the future.
Georgia Tech-Based Institute looks to the future
The future is also where Georgia Tech’s innovative team is focused. The university that brought Jill Watson to life has since rolled out two new AI tutoring tools: AskJill and Agent Smith. The Atlanta-based university also recently helped establish the National AI Institute for Adult Learning and Online Education (AI-ALOE) with a five-year, $20 million grant. years of the National Science Foundation.
The grant will fund “large-scale investments” in technology infrastructure, according to Ashok Goel, professor of computer science at Georgia Tech and executive director of AI-ALOE. Investments will go to data storage, compliance and cybersecurity.
AI-ALOE is particularly interested in adult learners, as the name suggests. Goel says he predicts millions of Americans will need to be re-skilled or up-skilled over the next decade because automation will change the way we work, and those people won’t be traditional students. They’ll have families, jobs, and other responsibilities, and they won’t be able to attend their classes until the wee hours of the night or on weekends, when professors, teaching assistants, and human tutors aren’t around. available, increasing the need for AI assistance.
READ MORE: Improve online learning and more with artificial intelligence.
As for the initiative itself, Goel wants to bring AI Tutors to the world at scale. Jill Watson is great, he says, but the effort it takes to create a Jill Watson makes it nearly impossible to replicate in other higher education institutions or K-12 environments. And the program’s other goal – improving the capabilities of AI itself – will go hand-in-hand with the expanded use of a tool like Jill Watson.
Goel says Georgia Tech’s AI tools have learned from more than 40,000 user questions over the years. It might sound overwhelming at first, but Goel says 4 million questions would be a much nicer base to learn from.
Even though AI is getting smarter, the latest version of AI is designed to help extend its reach. Agent Smith (named after the autocloned antagonist in The matrix movies) offers the most intriguing potential for large-scale growth. Agent Smith is a cloner, capable of replicating a Jill Watson AI Tutor for courses and classrooms across the country in as little as five hours. It’s still too long, Goel says, and the AI tutor interface is still a bit clunky, but the team is making progress.
“Why can’t we bring AI tutors to every teacher, every learner, and every classroom in the world?” Goel asks. “If Jill Watson only remains the provenance of 30 classes, that’s interesting, but it’s not a game-changer. It’s only a game-changer if anyone can use it.