Written by Michael Plocher
For some time, artificial intelligence (AI) methods have been used in education. Until recently, most uses of AI focused on data analysis and have been learner-oriented, instructor-oriented, and institutional system-oriented (Zawacki-Richter, Marín, Bond, & Gouverneur, 2019). AI used in this way includes adaptive or personalized learning systems, such as plagiarism detection, automatic feedback on submitted work, and online quizzes that automatically adjust the difficulty of questions based on students’ responses.
ChatGPT – A Natural Language Processing AI
Within the last four years, a type of AI called Large Language Models (LLM) has enabled large-scale Natural Language Processing (NLP) research (Gruetzemacher, 2022). This form of artificial intelligence focuses on how computers can process language like humans do.
Barry (2019) of OpenAI blogged in February of that year about their NLP model, ChatGPT:
“We’ve trained a large-scale unsupervised language model which generates coherent paragraphs of text, achieves state-of-the-art performance on many language modeling benchmarks, and performs rudimentary reading comprehension, machine translation, question answering, and summarization—all without task-specific training.”
Instead of just looking at data and making decisions, OpenAI’s ChatGPT was trained to accurately predict the next word in a sentence.
This announcement created a buzz among educators who began to wonder how and when this would impact their teaching. Could they use it to plan lessons better? What about having it assist with developing courses? Developing assignments? Would it help students cheat by writing papers?
Educators didn’t have to wait long to get access to ChatGPT. At the end of November 2022, OpenAI released a free version of ChatGPT and, soon after, a paid version. What follows in this article is current as of March 2023; however, this area of AI continues to change at a staggering pace!
Other AI Sites
Programmers can now incorporate an AI like ChatGPT into their own projects and websites. Enter some lines of text and these websites can generate an essay, create test questions, and summarize novels. They can generate music, art, polls, word clouds, and interactive presentation slides, and even turn your text into voiceovers spoken by real human-looking avatars.
Microsoft began incorporating ChatGPT in its Bing search engine in February this year. In March, Google released their own version called Bard, while Apple announced they are exploring something similar yet have nothing to show.
How Does It Work?
To effectively use an AI like ChatGPT, you need to prep it well to get the results you want. This means providing an explicit prompt, giving it a role or voice, and setting parameters of your answer (Fitzpatrick, Fox, & Weinstein, 2023). Even when well prepped, however, the AI responses might be incorrect. You still need to check them for biases and misinformation.
What Does This Mean for Teaching?
AI can now write lessons plans, computer code, research reports, persuasive essays, and book reports, and can generate math problems with answers and explanations, etc. In the classroom an activity can include assigning students to have an NLP write a paper for them. Students would then spend a few class periods analyzing and evaluating the text, working on reading comprehension skills (Klein, 2023), finding sources, rewriting portions of it to make it match their own style, and filling it with their own “emotion,” as one of my middle school students told me. These analyzing, evaluating, and creating skills are higher-order thinking skills that we want our students to develop.
AI can also be used to help educators reduce their workload (Baker & Smith, 2019). It can generate the text for a lesson, a human avatar that can speak the lesson, presentation slides with images, and assessments that can be automatically assigned and graded. Using AI to assist in this way, however, creates many other questions that need to be explored further.
Literacy Skills Are Still Needed
The ability of AI to generate papers and provide other responses that are considered forms of summative assessments has led some schools to ban its use among students. Their argument is that using it will lower literacy skills and will allow students to cheat more easily. However, NLPs need an input of concise, clear language to return the correct information. If someone with poor literacy skills provides input, it can lead to an output that has incorrect information, grammatical errors, and spelling and syntax mistakes. This means educators will still need to ensure students have a high level of literacy skills (Fitzpatrick, 2023). These skills will be increasingly important to successfully interact with AI.
Going Forward
With the release of ChatGPT, many feel this is the true beginning of the age of AI since we now can have advanced human-to-machine communication. As NLPs advance, allowing improved spoken functionality, it will allow Conversational AI technology to use NLP AI for more personalized, accurate, and natural responses. We are getting closer to a goal of education in which teachers using AI will quickly provide personalized instruction for each student, for at least part of every lesson (Seldon & Abidoye, 2018), ensuring each student’s God-given gifts and abilities are encouraged and strengthened.
Other than using AI to check grammar, spelling, and syntax, this article was typed and revised by the author—not AI.
Michael Plocher (BS in Ed DMLC ’93, MS Ed Tech MLC ’16) teaches at St. Paul Lutheran School-New Ulm MN, focusing on science, math, and ed tech. He also teaches the undergraduate course EDT3002 Teaching with Technology at Martin Luther College as an adjunct instructor. This June he will present on ChaptGPT/AI at the OpenLearning@MLC Conference.
References
Baker, T., & Smith, L. (2019, February). Educ-AI-tion Rebooted? Exploring the future of artificial intelligence in schools and colleges (Rep.). Retrieved March 27, 2023, from Nesta website: https://media.nesta.org.uk/documents/Future_of_AI_and_education_v5_WEB.pdf
Barry, B. (19, February 14). Better language models and their implications. Retrieved March 27, 2023, from https://openai.com/research/better-language-models
Fitzpatrick, D. (2023). Unlocking the Power of AI: Why Literacy Skills are Key to Success in the AI Era [Web log post]. Retrieved March 28, 2023, from https://teachergoals.com/literacy-skills-and-ai/?v=7516fd43adaa
Fitzpatrick, D., Fox, A., & Weinstein, B. (2023). The AI Classroom: The Ultimate Guide to Artificial Intelligence in Education. Beech Grove, IN: TeacherGoals Publishing, LLC.
Gruetzemacher, R. (2022, April 22). The power of Natural Language Processing. Retrieved March 27, 2023, from https://hbr.org/2022/04/the-power-of-natural-language-processing
Klein, A. (2023, March 15). Measuring Reading Comprehension Is Hard. Can AI and Adaptive Tools Help?. Retrieved March 30, 2023, from EducationWeek website: https://www.edweek.org/technology/measuring-reading-comprehension-is-hard-can-ai-and-adaptive-tools-help/2023/03
Seldon, A., & Abidoye, O. (2018). The Fourth Education Revolution: Will Artificial Intelligence liberate or infantilise humanity. Buckingham, United Kingdom: The University of Buckingham Press.
Zawacki-Richter, O., Marín, V. I., Bond, M., & Gouverneur, F. (2019). Systematic review of research on Artificial Intelligence Applications in higher education – where are the educators? International Journal of Educational Technology in Higher Education, 16(1). doi:10.1186/s41239-019-0171-0