Each school year presents a unique set of challenges to prepare students for future success. This year, machine learning, or artificial intelligence (AI) as it is more commonly known, has become a prioritized area of focus for educators at every level of the school system — and for good reason. In a world where nearly every aspect of communal life is driven by data and technology, educators are recognizing the transformative potential of AI to address many longstanding challenges within education systems.

One such enthusiastic educator is Sal Khan, founder and CEO of Khan Academy, an online tutoring platform that has been widely used and promoted in classrooms around the world for nearly two decades. In a recent TED Talk, Khan explained why the nonprofit made the choice to use generative AI technology to enhance the user experience for both teachers and students.

The impetus to add an AI-powered assistant that acts as both tutor and teacher assistant to the platform came about from a concern for the potential for continued learning loss and an understanding of human nature. Because generative AI for public use is part of open educational resource (OER) technology that is still very much in its infancy, the Khan development team understood that much of the programming in the public domain accessible to students is of poor quality and that, naturally, students may be tempted to use these OER programs to complete assignments or get answers quickly.

But because the learning process is eliminated from the AI transaction, students may not have the skills needed to recognize the poor quality of the content without proper guidance and support. With thoughtful programming, these OERs can restore the learning process and deepen learning in many ways.

Like the Khan Academy developers, many educators in the field are banding together in grade-level teams, and more widely through social media applications, to share strategies for harnessing the power of generative AI both as a learning tool and as a teaching assistant. The classification “generative” comes from this type of programming’s ability to generate content when prompted through human input in the form of a question or command. Generative AI can produce a variety of content through the recognition of patterns in language and images drawn from a large language model (LLM). These LLMs, which begin as neural networks, are fed content from the world wide web and other sources to create large data sets that engineers then train to produce human-like output using algorithms.

It is exactly these anthropomorphic characteristics of the output that can draw concerns around learning loss and plagiarism. AI is now embedded in every sector of modern life and LLMs are not going away, and will only improve in their ability to mimic humans in music, art and written works.

To introduce AI as an additional tool for completing tasks to demonstrate competency and comprehension, educators are also exploring the use of plug-ins — accessory programs that can automate the more rudimentary tasks associated with developing instructional and learning sequences — and mastery of the critical skill of prompt engineering, which is structuring a prompt input to narrow output. These two strategies can help educators to maintain instructional control, tailor instructional delivery to individual student need, and train selected applications to elicit responses to student-led queries that are more ideally Socratic.

One of the most significant advantages of bringing AI into the classroom is the potential it poses for allowing teachers to determine baselines more readily and then craft and deliver tailored learning experiences to more students. One of the most cited reasons leading to educator burnout is the feeling at the end of each day that someone was left behind. And with rising student-to-teacher ratios and necessary schoolwide instructional pacing, many students for whom personalized attention could make a world of difference, often are left behind. AI may serve as a tool to minimize gaps in learning through an increased capacity for differentiating the curriculum according to individual need.

The world has transitioned from an industrial-based society to one based on service to the current knowledge-based economy. And never has the imperative for preparing students for this next reality been so great. Today’s student will need an additional skill set beyond traditional learning targets to keep pace with rapid changes in technology and compete in future job markets developing from this next frontier in technology.

To help address some of the ethical and procedural concerns around the use of AI as an educational tool in schools, CSBA has updated the following sample policies, available in the September 2023 Update Packet:

• BP 5131.9 – Academic Honesty
• BP 6154 – Homework/Make Up Work
• BP 6162.5 – Student Assessment

Future policy revisions will likely follow as it becomes evident other ways AI will impact education.