I recently had a discussion with a friend of mine about the future of instruction and learning post ChatGPT. Not sure that we had covered all the possible ways that teachers could or should modify their instruction, I decided to pose a question or two to ChatGPT4. Below are the questions/prompts and the responses.
With LLMs like ChatGPT able to answer questions in essay form, how should teachers restructure their classes to enable students to learn without consulting the LLM?
The rise of sophisticated language learning models (LLMs) like ChatGPT poses a unique set of challenges and opportunities for educators. While these AI models can be a fantastic resource for information, they may prompt students to take shortcuts in their learning, perhaps relying on the AI to do the heavy lifting rather than deepening their own understanding. In response to this, teachers need to rethink their instructional strategies to better engage students and motivate independent learning. Here are a few strategies that teachers might consider:
- **Focus on Critical Thinking and Interpretation**: Encourage students to think critically and interpret the information they access. Teachers can design activities that require students to apply, analyze, and evaluate the knowledge they gain, rather than just recall facts. This might involve tasks like discussing the implications of a historical event, designing an experiment to test a hypothesis, or evaluating differing viewpoints on a complex issue.
- **Emphasize Process over Product**: Focus on the learning process as opposed to just the final product. When assigning tasks or projects, ask students to document their thought processes, research steps, and decision-making along the way. This helps students learn to navigate the learning process independently, rather than just presenting a polished final product that might have been heavily influenced by an LLM.
- **Promote Creativity and Originality**: Assignments that require original thought, creativity, and innovation are less susceptible to AI shortcuts. For example, writing an original short story or creating an art project can’t easily be done by an AI. This kind of work encourages students to think independently and to apply and synthesize what they’ve learned in new and unique ways.
- **Implement Collaborative Learning**: Encourage group work and peer-to-peer learning. By working together, students learn to communicate effectively, to negotiate differing viewpoints, and to collaboratively construct knowledge, skills that an AI can’t easily replace.
- **Teach Digital Literacy**: It’s also important to teach students about the limitations of AI and to develop their digital literacy skills. This includes teaching them to evaluate the quality of the information they access online and to understand how AI can be both a tool and a crutch in their learning.
- **Encourage Inquiry-based Learning**: Rather than providing direct answers, teachers can guide students towards asking their own questions and seeking answers independently. This develops skills in research, problem-solving, and self-directed learning.
- **Real-world Application**: Assign projects or tasks that require students to apply their knowledge in real-world scenarios. This could be community-based projects, service-learning opportunities, or just practical applications of the concepts they’re studying.
The key here is to emphasize the value of learning and independent thought over simply acquiring information. As educators, it’s our job to help students become lifelong learners, and in a world increasingly filled with AI, this skill is more important than ever.
If I provide students with a reading assignment about an historic event and ask them to comment on the implications of that event, how do I know that they didn’t input the text into an LLM and ask it to provide the implications of the event?
It is a valid concern and there are few methods you could implement to ensure that students are doing their own thinking rather than relying on an AI model like ChatGPT. Here are some strategies to ensure authentic student responses:
- **Design Complex, Multi-Faceted Assignments**: When assignments require students to draw upon multiple sources or incorporate their personal viewpoints, it becomes more difficult for AI to generate an appropriate response. For instance, you might ask students to compare the historical event with a current event or personal experience.
- **Promote Reflection**: Ask students to provide reflections on their learning process and personal connections. A reflective component could be a powerful tool to ensure the originality of their work. It’s much harder for an AI to fabricate a student’s personal reflections or learning journey.
- **Scaffold the Learning Process**: Break the assignment into smaller parts and have regular check-ins. For instance, you could first ask them to summarize the event, then later to analyze the causes, and finally discuss the implications. This also allows you to see their thought process unfold over time and provide guidance along the way.
- **In-Class Discussions**: Incorporating class discussions can help you gauge whether students have genuinely understood the material. This could take the form of small group discussions, whole class debates, or individual presentations. This also provides an opportunity for peer learning.
- **Utilize Plagiarism Detection Software**: While it won’t directly detect AI-generated text, these tools can alert you to unusual patterns or phrasing that might indicate use of an AI.
- **Teach and Uphold Academic Integrity**: Make sure that students understand what constitutes plagiarism and the consequences of such actions. Explain that using an AI to generate their work is a form of academic dishonesty. Make it clear that the goal of education is not just to complete assignments, but to gain skills and knowledge that will be valuable throughout their lives.
- **Explicit Instruction about AI Tools**: Provide explicit instruction about the ethical use of AI tools. Discuss the difference between using AI to support learning (e.g., to explain a difficult concept) and using it to substitute their own thinking (e.g., generating an essay).
Keep in mind that the goal isn’t to catch students cheating, but to encourage genuine understanding and independent thinking. These strategies can help ensure that students are actually engaging with the material and learning from the assignments they complete.
ChatGPT provided two interesting responses to my prompts (neither response was edited). When I think about this holistically, I believe the prevalence of LLMs will force more instructors to collaborate with their colleagues regarding tactics for utilizing different assignments and tests that they have utilized in the past. I have read some opinions of academics who are calling for more oral examinations. While that may be possible for in-person classes, it’s more difficult to do for online courses and their students. It’s also likely more expensive in the long run.
It’s too soon to predict the impact of these LLM tools on education in general. Those who are attempting to stave off the impact by banning these tools are sticking their heads in the sand. The last paragraph of the response to the second prompt is an appropriate admonition to them – “Keep in mind that the goal isn’t to catch students cheating, but to encourage genuine understanding and independent thinking. These strategies can help ensure that students are actually engaging with the material and learning from the assignments they complete.”
After all, learning was the future and continues to be the future for mankind. We just have to craft solutions to ensure our students learn enough to be able to contribute to their self-preservation, their community, and society in general.