Ellen C. Agnello

AI. It’s More than ChatGPT

The recent surge in Artificial Intelligence (AI)-powered applications over the past year have elevated AI to conversation king, leaving many, including teachers and school leaders wondering how will AI and ChatGPT impact K12 schools. According to a survey by Pew Research, most Americans are equally concerned and excited by these new advances, but less than a third of Americans are highly aware of the fact that they’re already using AI in their daily lives.

AI refers to applications in which computers are programmed to perform a variety of human cognitive tasks (Nilsson, 1998). These include chatbots that answer customer questions, wearable fitness trackers that analyze activity and sleep, product recommendations based on shopping activity, and an email host categorizing a message as spam.

What is AIEd?

AI in Education (AIEd) refers to these technologies applied to the school or district setting. While the release of ChatGPT in November, 2022 has ignited much debate and polarization among K-12 education stakeholders, it is important to note that AIEd has a history and cannot be reduced to a single application.

For instance, computer-assisted spell and grammar checkers were developed over sixty years ago. In 1959, a research team from the University of Pennsylvania’s Department of Linguistics created the first computer program to identify the case, tense, and function of a given verb. In 1971, Ralph Gorin, a graduate student at Stanford University’s Artificial Intelligence Laboratory, created the first spell checker application program, though others had previously been developed for research purposes. Ten years later, in 1981, Drs. James and Janet Baker developed an application that could recognize single, spoken words. They released Naturally Speaking a decade later, which became the first continuous dictation product and was the basis for programs including LG Smart TVs voice recognition feature and DragonDictate for Mac. These AI applications have been used in schools for decades without sounding alarm bells or inciting frenzy.

Perceptions of AIEd

AI has been used to enhance user experience in applications in many industries. It’s used by Google Maps to identify and forecast traffic slowdowns, by Amazon to recommend products to shoppers, and by Apple iPhone to flag a call as potential spam. While many of these AI applications have been welcomed by consumers who appreciate the conveniences they provide, the thought of integrating similar technologies into the school setting is less palatable for many. A recent survey of nearly 1,000 teachers, principals, and district leaders by Edweek’s Research Center found that about half of respondents predict AI will have a negative effect on teaching and learning. Less than one-third reported that they think it will have a positive effect. Interestingly, 14% of respondents reported that they don’t know what AI platforms are.

Perhaps the majority-negative sentiment towards AIEd stems from a lack of understanding or from misunderstanding it to refer only to natural language processing technologies, like ChatGPT. The findings of a content analysis of research studies on AIEd by Guan et al., 2023 found that three main categories of AI applications persist and are used in schools already (Guan et al., 2023). These are:

1. Learner-oriented AIEd which include programs, like the personal tutor, Amy. These enable students to independently study a subject domain with the help of a chatbot expert;
2. Teacher-oriented AIEd which include programs, like Otter.ai which automatically transcribes meeting minutes and could be useful during PPT or PLC meetings. These applications automate administrative tasks related to the job like grading, assessment, detecting plagiarism, providing feedback, progress monitoring, and risk identification, and;
3. School/District system-oriented AIEd which provide insights about broader statistical patterns like those related to student enrollment, dropout, graduation, etc. See how the Uruguayan government is employing AI to predict dropout here.

How Has AI Impacted Schools Already? What Research Says

While this field is relatively new within the research context, a content analysis of 425 published AIEd articles in peer-reviewed journals between 2000 and 2020 revealed the following four categories of interest, which suggests that these are the most common applications of AI in education during that twenty-year period.

Adaptive systems and customization

62.6% of articles explored AI technologies in this category. These include AI computer-assisted instruction, virtual and augmented reality, educational games, and adaptive teaching and learning.

Intelligence tutoring systems (ITSs)

9% explored technologies including predictive modeling and assessment design.

Assessment and evaluation

20.9% of articles explored technologies in this category. These include computer programs that provide immediate and tailored instruction or feedback to students, typically without human instructor intervention.

Profiling and forecasting

3.8% explored models that enable the active tracking of student attributes like knowledge, engagement, academic performance, and/or risk of dropping out.

How Will ChatGPT Impact Schools?

ChatGPT by OpenAI is a large language model (LLM) that was launched in November, 2022. By January 2022 it had over 100 million active users. But where does it fit into the current AIEd categories?

Learner-oriented: The most obvious use cases is to search for information and to produce writing. But are these the best uses? Many worry that students will fail to develop important skills if they rely too heavily on the technology.

Teacher-oriented: ChatGPT can write lesson plans and learning objectives, it can recommend materials to teach curricular concepts, and it can even simplify text (see below).

School/District system-oriented: Schools and districts can use it to make communications more accessible to families for instance, by having it translate notices, report card comments, permission slips, etc., into different languages (see below).

Time will tell how this and other AI technologies will shape K12 teaching and learning.

References

Guan, C., Mou, J., & Jiang, Z. (2023). Artificial intelligence innovation in education: A twenty-year data-driven historical analysis. International Journal of Innovation Studies, 4(4), 134-147.