Artificial intelligence technology can tell doctors when a scan reveals a tumor, can help the military distinguish between a truck and a school bus as a target, and can answer a high volume of college students' questions.
Sectors of our economy such as the military, health care, and higher education are much further along than the K-12 system in incorporating artificial intelligence systems and machine learning into their operations. And many of those uses-even when they are not specifically for education-can spark ideas for applications in K-12 that may be more pertinent than ever imagined.
With the coronavirus upending traditional ways of delivering education, AI technologies-which are designed to model human intelligence and solve complex problems-may be able to help with logistical challenges such as busing and classroom social distancing, provide support to overwhelmed teachers, and glean new information about remote learning.
AI techniques and systems are "like the internal combustion engine-you can use them to power a lot of different things," said David Danks, a professor of philosophy and psychology at Carnegie Mellon University in Pittsburgh, who studies cognitive science, machine learning, and how AI affects people. "The exact same thing can be used to predict whether someone has cancer, or whether students understand a concept, or to classify somebody as a bad guy you want to go after."
Of course, there are lots of potential trouble spots when thinking about the role of AI in K-12 education. Artificial intelligence learns from the data that are fed into it, and if that input includes bad data or data applied incorrectly, poor or biased decisions may result. At the same time, the use of AI in K-12 raises very serious data-privacy concerns because such technologies would likely be used to personalize education or make important decisions for individual students.
But even with those concerns, AI advocates say other sectors are already offering lessons learned for how the technologies could be used in K-12 for teaching and learning and the management of schools. That is especially the case with the military, health-care, and higher education fields.
Here is a look at what K-12 educators, policymakers, and planners could learn from those three sectors:
"Nearly every military in the world believes that advances in AI will play a critical role in shaping the future of military power. But there are big disagreements about what is possible and what is wise." -Michael C. Horowitz, a professor of political science who studies military innovation at the University of Pennsylvania.
Military leaders are using AI simulations to assess military tactics and determine the likely outcome of strategic plans. Plugging different variables into these scenarios-everything from weather predictions to the timing of attacks and estimating troop numbers-can show how outcomes might change. Also, soldiers can get important practice in simulated real-world settings with low risk.
> K-12 Applications: AI-powered simulations could be useful for planning purposes for everything from scheduling to determining the most effective models for social distancing when students return to their school buildings amid the COVID-19 outbreak. Some companies are already using simulations to train educators on successful techniques to help students with social-emotional learning, trauma, and mental-health issues.
Tanks, airplanes, submarines, trucks-all that military equipment needs to be maintained to keep troops safe and operations running smoothly. Some high-tech AI systems can predict when parts need replacing before they break or when systems need tuneup. Artificial intelligence has helped the military optimize in-flight refueling of jets to make the dangerous technique safer and more efficient.
> K-12 Applications: School districts also rely on a lot of equipment-think buses, computers, air conditioning systems, and more. AI-powered smart programs are already being used in some schools to fine-tune building operations, lower energy costs, and manage maintenance and repairs.
The backbone of the military revolves around logistics and supply-chain management. How to get equipment and personnel from point A to point B most efficiently and cost effectively is something that AI systems are tackling for the military.
> K-12 Applications: The uses are widespread: AI systems could optimize scheduling, the distribution of laptops, cafeteria operations, and bus routes. In fact, the Boston school district has saved more than $5 million using a high-tech AI system that streamlined bus routes.
"At its core, AI is really about using big data to be able to help predict what will happen so we can show up at the right time with the right solution." -Ron Goldman, the CEO and co-founder of Kognito, a company that uses an artificial-intelligence-powered platform primarily for mental-health training in the health-care and education fields.
Artificially intelligent technologies can analyze radiology and CT scans looking for abnormalities. Programs can quickly sift through images much faster than humans and identify patterns based on vast data. These techniques can identify tumors and health issues and suggest treatments, which are then reviewed by medical professionals.
> K-12 Applications: Programs powered by artificial intelligence could do a better job identifying student risk factors and recommending earlier and more targeted academic or mental-health interventions. The goal isn't to replace teacher decisions but to save teachers time and to amplify their own expertise. Using big data and AI to spot patterns might be applied to other situations, such as taking student temperatures to check for COVID-19 before they enter school buildings or being able to target outbreaks more quickly.
Access to massive amounts of digital medical data and the use of AI to analyze it are making it easier to personalize medical treatments for patients. AI can predict how someone's current health behaviors are likely to affect their future health outcomes. High-tech systems can design much more sophisticated drug and treatment strategies tailored to an individual patient's biology or type of cancer, for example.
> K-12 Applications: Many education companies already talk about being able to help personalize the learning experience for students, but this is still just an emerging effort in most places. Some K-12 programs are using artificial intelligence to collect data on student behavior and academic engagement and then guide students through suggested individualized lessons. CENTURY Tech, a London-based company, for example, uses an AI platform that tracks student interactions and behavior patterns and academic performance to create personalized learning paths.
Artificial-intelligence-powered programs are being used to train medical professionals in many ways. AI company Kognito, for example, uses its health simulations to help doctors and nurses practice discussing and interacting with patients around sensitive topics like obesity, mental health, and suicide. Through conversations with "virtual humans," medical practitioners can practice and model effective techniques.
> K-12 Applications: Kognito has a version of its product that is designed for educators, training them to lead conversations with students around social-emotional learning and mental health, using research-based language and techniques. An expanded version of this technology could be applied in other areas. About 15,000 K-12 schools currently have access to Kognito simulations.
Early medical intervention, making sure patients adhere to treatment, and supply-chain management are all ways that AI can affect the bottom line in various aspects of health care.
> K-12 Applications: The same goes for schools and districts. AI-powered programs could predict what supplies are needed and where with more accuracy, analyze budget trends, and identify spending patterns in areas ripe for savings, especially given that K-12 budgets are likely to be slashed significantly as the economy struggles through COVID-19.
"There's not an obvious wall between higher education and K-12 [around uses for AI]." -Jacob Whitehill, an assistant professor of computer science and a core faculty member in the learning science and technology program at the Worcester Polytechnic Institute in Massachusetts.
What if teachers could have more information in real time about whether their students grasp concepts or are struggling when learning online? Whitehill is exploring the idea of an AI-based program that uses a video camera to take many small snapshots of students as they learn remotely to analyze their facial reactions. Such a program would provide teachers with real-time feedback on students' cognitive and emotional states. (But that program is also just the kind of technological approach that would prompt intense criticism from student-data-privacy advocates.)
Virtual Teacher's Assistant
When Georgia Tech interactive-computing professor Ashok Goel was having a hard time answering all the questions coming from the hundreds of students in his online computer science class, he created an artificially intelligent tutor he dubbed Jill Watson. "She" was able to answer many of the students' more routine questions, freeing up time for Goel to do higher-level work. Since that first experiment, Watson is now used in 17 online classes, Goel said, covering more than a thousand person-hours of work. Goel, who is also the chief scientist for C21U, a company developing innovative uses for AI, is now working to adapt Watson for high school and middle school teachers. And with remote learning, he believes the AI teaching assistant could also be used to help answer parents' questions as they support students at home.
Colleges and universities are already using this approach to some degree, and this latest version of that technology is moving into the K-12 education space. Automated AI essay graders have been around for some time, but the makers of the software say the AI features now available are much more sophisticated evaluators of student writing than what were available years ago. They can judge hundreds of features in a written piece, everything from spelling and grammar to sentence structure. (Lots of concerns remain that these programs can be biased, can fail to interpret creativity correctly, and can be "gamed" by students writing to the algorithm.) Though some states are using these types of programs to grade essays on their standardized state tests, they're yet to be widely adopted on a district and school level.