AI in Education: What could it look like?
The future is here and it's only just beginning. Here’s our initial take of what’s happening with AI in education.
I think it’s safe to say that advances in artificial intelligence (AI) is a hot topic in education news. David Lumb from CNET described how the AI tool ChatGPT from OpenAI can be used to save you money, build a weight loss plan, and even write your grocery lists for you. Nicole Krueger recently wrote why students should use AI tools for creation. Greg Toppo from The 74 adds that ChatGPT can also take pretty solid notes and summarize a video or audio file for a user. All of these examples go a long way to helping students with different learning styles and learning needs and has the potential to help students become producers of content rather than simply consuming knowledge. The use of tools like ChatGPT in education also has potential negative consequences if we don’t evolve our methods of assessment of learning. For example, New York City schools has decided to ban ChatGPT over fears of “negative impacts on student learning, and concerns regarding the safety and accuracy of content.” Accommodating these rapidly advancing technologies will be both an opportunity and a challenge for educators and school leaders in the years ahead. Even the Office of Educational Technology of the federal Department of Education has published a series of papers about the AI in education.
Regardless of whether you’re all in or all out on AI in education, it’s important to be aware of this technology and how it might affect teaching and learning in the years ahead. Here are some possibilities we’ve just begun to explore, and continue to research as part of our work matching promising technologies with the needs of education.
AI Enables Students to Focus on Application, Insight, and Creativity
Contributed by Jonathan Maier
By and large, electronic calculators are accepted tools in today's mathematics classroom. The rationale goes like this: “let's not train students to be number-crunching machines, when machines are readily available to do the work, and they're really good at it. Let’s instead have students think about the mathematical concepts behind the calculations.” But what if machines are not only good at tedious computations, but become highly skilled in much of what we ask students to do at school: write an essay, solve a story problem, make a chart, conduct research, draw a picture, program software, or engineer a robot?
Artificially intelligent machines still have a ways to go to reach that bar. AI’s generated responses to composition assignments have been described as "superficial" and "that of an earnest 8th grader.” They often lack the common cognitive processing mistakes made by learners (e.g., the conclusion doesn’t follow from the premise). And AI can be woefully short on common sense, though there’s significant work being done to rectify that (see this).
If AI tools improve sufficiently—and are sufficiently available—assignments and the work of students will change. Formulaic and impersonal tasks easily performed by AI, such as solving single solution story problems or essays adhering to a strict format, will give way to open-ended or highly personalized projects. Fortunately, this already mirrors some of the most potent teaching practices today. AI will take care of the mechanics, while the students provide the application, the insight, and the creativity.
And since AI excels at the kinds of rote tasks measured on standardized tests, in the future it could save students valuable time by taking those kinds of tests for them instead.
Affordable AI Leads to New Opportunities
Contributed by Willy Kjellstrom
The coming year will find machine learning models becoming more portable, less memory and compute intensive, and less reliant on Internet connectivity or cloud resources. Machine learning models—files that are trained to recognize patterns in data—are already being ported to smart devices that have computational capabilities beyond yesterday’s computers. However, advances in algorithm optimization have allowed people to implement AI on far simpler computers like microcontrollers using TinyML models. Because microcontrollers are often cheaper and smaller, deploying machine learning models on this type of edge device opens new possibilities for AI innovations in schools, communities, and the world at large.
Although most educators do not have expertise with machine learning or microcontrollers, new projects that incorporate TinyML frameworks on low-cost, low-power are being developed all of the time. It’s not inconceivable to speculate that schools could spend less than $50 on a device that is trained to do the following:
- Monitor bird species around a school bird feeder;
- Securely count and identify students entering a building, school bus, or classroom without connecting to the Internet (or taking time to take attendance each day);
- Give feedback to teachers about how much time they are spending on direct whole-class instruction instead of individualized or small-group instruction;
- Identify optimal placement of school crossing guards based on traffic patterns; and
- Make recommendations about which students might need more instructional attention based on how quickly they move through rotation stations in your classroom.
Imagine the rich, place-based learning opportunities that might occur when AI becomes a part of a curricular project or course of study.
Calculator is to Math as AI is to Other Subjects
Contributed by Tim Rayle
Generative AI has been compared to inventions of the camera or the calculator. Did the camera end the need for art skills? (No.) Did the calculator end the need for math skills? (No again.) These technologies certainly brought about changes to the math and art worlds, and generative AI will bring changes as well. These tools are here, they’re improving and are already changing education. Just as math education shifted from a focus on algorithmic processing to the underlying mathematical concepts with the widespread adoption of the calculator, AI may be the catalyst for other subjects to shift to a focus on deeper understanding and comprehension by learners not easily reproduced by a generative AI.
English teachers will have to adapt the same way that math teachers have adapted to the calculator. There will have to be some assignments where ChatGPT is permitted, even encouraged as a tool and others where it’s discouraged. Perhaps this means more oral, video, and other kinds of in-class assignments and assessments? Perhaps AI could be used as a tool to get beyond writer's block, to produce outlines and first drafts, with refining and editing done by the student. Educators will need to teach students to use tools like ChatGPT to help them effectively express the ideas that they want to communicate.
Our Focus Is on the Future
At Clarity Innovations, we imagine, design, and build the future of education. Much of what we do is look at emerging technologies to find ways they can be used to improve the process and practice of teaching and learning. AI is just one of those that shows tremendous promise (as well as uncertainty). If you’re interested in learning more about what we see for K-12 education 3-5 years from now, please contact us. We’re here to help you succeed.