What will artificial intelligence (AI) eventually look like in our classrooms? Ethan Mollick, professor at the Wharton School of the University of Pennsylvania and cofounder of Wharton Interactive, predicts that AI will be undetectable, ubiquitous, and transformative. With the rate that AI has fueled new tech start-ups and is being integrated into businesses, this seems very reasonable.
While we may not be there quite yet, it is not too early to consider the ways in which we can begin integrating AI into our classrooms and, when possible, have students learn how to effectively use these transformational tools. After all, students are our future change agents, and we need to empower them for their rapidly evolving world.
If Students Cannot Access AI Tools
If your students are too young or aren’t allowed to access generative AI tools for any reason, you might consider modeling its use as the classroom teacher. This modeling can allow you to elicit input from your students and have them collaborate in your demonstration of the tools.
If Students Can Access AI Tools
If your students are allowed and able to access generative AI tools, there are extensive ways in which you can have them interact with and leverage these programs to increase productivity while also learning the subtle nuances involved.
Ethan Mollick says, “But there is no operating manual for LLMs, you can’t go to the world’s top consultancies and ask them how to best use LLMs in your organization—no one has any rulebook, we are all learning by experimenting.” With this context in mind, it’s highly beneficial for our students to have a chance to experiment with AI tools. It can provide them with an opportunity to begin mastering the applications that will likely transform much of our future world. In fact, in his college courses, Mollick actually requires his students to use AI, and he has found that it has allowed his students to excel.
Potential Student Uses of Generative AI
The following is a list of possible student uses for generative AI. This is not intended to be an all-inclusive list, and you may find that some of these uses don’t meet your specific needs, which is perfectly fine. The list is intended to be a conversation starter and a place for you to begin thinking about how your students, when allowed, might use generative AI for learning in your classroom:
Students can paste in text and ask the chatbot to translate it to another language.
Students can paste in articles, notes, or anything text-based and ask the AI to summarize the content. This can help students synthesize and process large amounts of information.
Students can paste in text and ask the AI bot to rewrite it at a different reading level to improve accessibility for struggling readers.
Students can ask the bot for help generating ideas around a topic. This can help initiate a brainstorming process and prompt other ideas from the students. Students will still need to review any suggestions and make judgments about their value.
Students can input ideas or summarize a topic, and then ask a chatbot for possible ways to outline this information. As with brainstorming, students will need to review and refine these starting points and make judgments about what will work for them.
Much like we use spellcheckers and grammar checkers that are integrated into our word processing programs, students can utilize generative AI chatbots for this purpose. Students can paste in their writing and ask the chatbot to proofread for specific types of errors—such as punctuation, grammar, spelling, and formatting—and suggest improvements. Students should then consider each suggestion and determine if they wish to make the change or not. Rather than requesting the AI to proofread in general, it’s usually best to identify one proofreading task at a time.
Similar to proofreading, students can ask a chatbot to review writing for things like sentence structure, clarity, and support material. While chatbots are better at proofreading than editing, they can also provide insights in this area. Again, it’s usually best to ask it to complete one specific task at a time. Students might even get as specific as, “What are some other ways I could restructure this sentence?” Apps created specifically for the editing process, such as Grammarly, can also be leveraged.
Students can ask the chatbot to provide background information on almost any topic. Because AI chatbots usually retain the context developed in a string of queries, students can ask follow-up questions to learn more and refine the information they are receiving. This can be a great way to gain a general understanding of a topic, which can inform more targeted research.
Students can ask the chatbot to guide them through the writing process. For example, they could tell the chatbot that they need to write a persuasive essay and would like the chatbot to guide them through the process. The student will still be generating their own ideas, but they’ll have a personal tutor or coach to guide them through the process.
Every day, new start-ups are developing applications built on open AI tools. For instance, the Merlin ChatGPT assistant is a Google Chrome extension that integrates generative AI functionality on websites. For instance, when doing a Google search, Merlin will provide a quick summary response next to your list of websites to review. It has the ability to summarize YouTube videos or articles, and it can be used as a writing assistant. Merlin users must be at least 13 years old.
Challenge your students to do something better than an AI chatbot. For instance, give groups of students a paper to proofread. Then, give the same paper to an AI chatbot. In comparing the work, glow and grow areas can be shared with students, and the class can engage in discussion around why there were differences between versions. This can be a great way to understand how a chatbot works while developing critical thinking skills.
Because a generative AI chatbot communicates with natural language, much like a conversation, students can ask it questions about what they are learning. In many ways, this is like having a personal tutor. Of course, any responses should be vetted for accuracy and bias, but this can offer immediate and helpful support. As a side note, products like Khanmigo from Khan Academy are currently under development, and these tools intend to integrate much of this tutoring capability and structure into the product. Until those products are widely available, these informal conversations with the chatbot can be effective.