To some, the thought of an AI-powered tutor might conjure up frightening images of a dystopian future, an iRobot scenario where near-sentient machines replace human beings in all walks of working life.
Indeed, people may be uncomfortable with the idea of artificial intelligence teachers and tutors because they serve as stand-ins for a profession that many believe you can’t — or shouldn’t — replace with robotics. These detractors would likely argue that AI tutors would quite literally take the humanity out of a job that requires human traits, like empathy and compassion, to foster effective education.
These naysayers, though, would be mistaken, and their concerns are often over-blown. AI tutors aren’t coming to replace teachers. Instead, they can be used to improve teaching practices and show instructors where they might be missing the mark with students in their sessions. When used effectively, AI can become a valuable tool for any educator, helping them by analyzing their teaching style, strategies, and a pupil’s general progress to show them the most optimal teaching strategies to enhance a student’s learning.
Such technology is already in use today; it’s not the practice of some far-off future.
One company, Third Space Learning, is using AI software to improve the teaching practices of its global network of 300 tutors. The UK-based company contracts out mathematic tutors from around the world to educate children from nearly 4,000 British primary schools.
In collaboration with scientists at University College London (UCL), Third Space Learning is trialing an AI tutor assistance. Developed from a data feed of tutorials that includes more than 100,000 hours of auditory and written information, the AI assistant is meant to provide instructors with feedback and guidance for their lesson plans.
The AI uses metrics for what it considers effective teaching practices to coach tutors on the best way to direct a tutorial, present information, or instruct a lesson. If, for instance, a tutor goes off-topic during a lesson, the AI might tell him to get back on track. Or say a tutor is talking too quickly and the pupil is losing interest or can’t keep up. In this case, the AI would tell the tutor to slow down so that her pupil can re-engage herself with the lesson.
As of now, the AI’s intervention will be minimal and confined to rudimentary improvements, but the organization hopes to scale it to offer more complex suggestions in the future. Even so, the program’s scientists and backers don’t think that the software, even in its most developed form, will serve as a substitute for teachers, but from their point of view, they do believe that AI has great potential to drastically improve current teaching practices.
“The potential for the use of AI to make education tractable and visible is huge,” Rose Luckin, a professor at UCL, told The Guardian. “What we are very interested in is the right blend of human and artificial intelligence in the classroom — identifying that sweet spot,” she continued.
“What I’m really concerned about is that people don’t run away with the idea that kids have to be plugged into the computer,” she said, conveying her fear that people will come away with the wrong impression for how the program improves teaching methods. “It’s about so much more than that.”
Tom Hooper, the CEO of Third Space Learning, shares her sentiment. “There’s a slightly dubious conversation about how AI will make humans irrelevant, but it’s not at all about replacing humans,” he iterated to The Guardian. “Our whole belief is that for children disengaged with the subject, who are lacking in confidence, people is what matter. An algorithm can’t provide that.”
For mainstream education, an algorithm cannot provide the human element that is so crucial to learning. But for secondary sources of information and less formal avenues of instruction, AI can play a more hands-on role. At least, Quizlet seems to think as much.
The online learning and educational resources company raised $20 mln in venture capital funding in February of this year to develop artificial intelligence services. The end game is to create an accessible, low-entry-cost alternative to private tutors for students all around the world who may not be able to access or afford additional educational assistance.
“This new functionality really helps give a guided study path that most effectively and efficiently gets someone to mastery on the information they’re studying,” Quizlet CEO Matthew Glotzbach told VentureBeat in an interview. “It’s that kind of investment — prior to tapping into some additional capital — we wouldn’t have been able to make. We just wouldn’t have [had] the resources to hire the caliber of people we need and invest proactively ahead of the curve,” he continued.
With the funding, Quizlet will expand its resources and staff to develop the product. As a 10 year-old company with plenty of online data to pull from — and now $20 mln to play with — the company likely won’t have any problem developing an AI model that can meet the the project’s expectations. As with Third Space Learning’s own project, however, Quizlet has no expectation to replace teachers or formal educational systems. Rather, it wants to provide an accessible and helpful resource to assist with the material that students are learning in the classroom; with AI, Quizlet is building a supplement, not a replacement, for in-school knowledge.
These are just a few of the companies exploring the use of AI and machine learning to improve education. Most of these companies leverage existing data for their AI models, and as such, they are limited to only the data they have at hand to enrich their services. Neuromation, on the other hand, employs synthetic data for machine learning analytics. With a platform like Neuromation, educational entities could speed up the rate at which their AI models learn, allowing them to create effective, life-changing AI tutors in a fraction of the time and at a fraction of the cost.
Using synthetic data to improve AI tutors is just one example of the numerous applications Neuromation’s platform can be used for. With Neuromation, the future is at our fingertips, and we’re just one synthetic data-enriched AI model away from taking hold of it for the better.