EDITORIAL: AI is an asset, but school must adapt

Like the calculator, computer and internet before it, artificial intelligence (AI) is quickly being embraced by students and teachers alike, many of whom turn to silicon-brained chatbots for all the mundane and repetitive tasks few are keen on performing. And just as those now-ubiquitous pieces of technology were initially opposed out of fears that classroom use would weaken students' fundamentals, AI is being pilloried for its potential to upset learning workflows and undermine scholastic achievement. That dismal scenario is by no means an inevitability, but schools must do all they can to ensure student success through safeguards and a clear-eyed approach that places AI firmly in the role of supplemental resource rather than replacement for critical thinking. Like the calculator, computer and internet before it, artificial intelligence (AI) is quickly being embraced by students and teachers alike, many of whom turn to silicon-brained chatbots for all the mundane and repetitive tasks few are keen on performing. And just as those now-ubiquitous pieces of technology were initially opposed out of fears that classroom use would weaken students' fundamentals, AI is being pilloried for its potential to upset learning workflows and undermine scholastic achievement. That dismal scenario is by no means an inevitability, but schools must do all they can to ensure student success through safeguards and a clear-eyed approach that places AI firmly in the role of supplemental resource rather than replacement for critical thinking. Kamehameha Schools is among the latest to officially onboard AI tools through OpenAI's ChatGPT for Teachers, a program that provides educators an institution-managed workspace to complete administrative labor. Functions include lesson planning, adapting instructional material and drafting emails, and educators receive training on how best to integrate those supports to increase efficiency. All fine — less time on banal administrative duties, more time for student enrichment — but the partnership and others like it fail to address student access head-on. ChatGPT for Teachers is, as its name implies, designed strictly for educators, and the Kamehameha Schools program does not provide student accounts. Perhaps it should, at some point. Children are early adopters and many are already experimenting with or using AI regularly. Monitoring and guiding student use of a new technology, be it calculators or AI, is much more difficult when curricula are blind to its existence or the tool is outright banned. With chatbots and similar resources just as capable of penning an 800-word essay as they are solving simple arithmetic, schools must steer toward ethical and transparent utilization from students. At this stage of deep consumer penetration, and ready access by anyone who owns an internet-connected device, prohibition is not an adequate solution. Human oversight is key. Without it, students might exhibit declines in critical thinking skills, the basic building blocks of analysis and evaluation that help students break down a problem before arriving at an informed answer. Communication is another issue. Studies have found that AI is reshaping human language, with conversational sentence structures morphing into truncated, stilted syntax formatted for machine digestion. And AI is far from infallible. Large language models such as ChatGPT are, at their core, prediction machines that use neural networks to regurgitate highly accurate — but not 100% — answers based on learned statistical patterns. No 'thinking' involved — yet. There is no right or wrong way to address AI — capabilities are rapidly evolving, making it a fast-moving target for educators — but there should at least be discourse and instruction in the classroom about how best to leverage it. The state Department of Education, which is dipping its toes into chatbot platforms, has rightly taken first steps by establishing general guidelines for both teachers and students covering a wide range of topics. Teachers are urged, for example, to put human inquiry first, be critical of AI content, cite generated content and regularly assess impacts on learning outcomes. Student guidelines include safety protocols and instructions to double-check AI content, be aware of bias and misinformation, provide citations for generated content and use tools only as instructed — no plagiarism allowed. A good start, but a broad push to more actively teach students how and when it is appropriate to tap AI, dependent on grade level, would be more advantageous. There is nothing inherently wrong with the adoption of new technology, particularly advances that increase efficiency and accelerate knowledge gathering. But AI is a different kind of beast, one so encompassing that it risks to degrade fundamental systems of thought should guardrails not be installed to curb its influence.