This post is part of a series on generative AI, informed by insights from my recently published book ChatGPT for Teachers: Mastering the Skill of Crafting Effective Prompts. Today, I want to highlight some key facts that educators and teachers need to keep in mind as they interact with AI in general. I choose to focus on ChatGPT for obvious reasons—it’s one of the most accessible and widely used AI tools in educational settings, making it a critical resource for enhancing teaching and learning experiences.
1. AI Enhances, not Replaces, the Role of Teachers
Chris Dede a senior Harvard researcher who spent years studying emerging technologies reported that when the first article about the use of AI in education was pubished in 1970, the author predicted that within five to six years AI will replace teachers becuase AI will be able to do almost everything.
Over 50 years now and we have yet to see a machine learning algorithm doing the task of teachers. The introduction of generative AI and especially large language models such as ChatGPT have and are still disrupting the notion of education as we have known it but is no where replacing the job of a teacher.
AI, I believe, can not replace teachers. it can compliment, facilitate, and enhance our teaching a machine remains a machine and teaching if an inherently social act where personal experiences and emotions play a critical role. The subtleties of human interaction, empathy, and the ability to inspire through personal stories are elements that AI cannot replicate.
This personal touch is what makes teaching not just a transfer of knowledge, but a transformative experience that shapes minds and hearts. While AI can provide tools and insights, the human element in teaching is irreplaceable, fostering connections and sparking curiosity in ways that technology alone cannot achieve.
2. ChatGPT Does not “Think” or “Feel” the Meaning of Words
When you interact with a conversational AI chatbot and provide a prompt, you receive an immediate response, creating the impression that the chatbot genuinely comprehends human language. While it does “understand” in a technical sense, it’s not in the way that humans do. AI chatbots perceive language as a collection of interrelated tokens—such as words, articles, and phrases—and make sense of these through complex algorithmic processes based on statistical probabilities.
Consider this illustration from the book ChatGPT for Teachers, which references an example from the Google Developers Blog:
When I hear rain on my roof, I _ in my kitchen.
To fill in the blank, a language model calculates the probabilities of various options, choosing the one with the highest likelihood. For instance:
- cook soup 9.4%
- warm up a kettle 5.2%
- cower 3.6%
- nap 2.5%
- relax 2.2%
In this scenario, the model selects “cook soup” as it has the highest probability. This example demonstrates that an AI chatbot’s understanding of language is fundamentally probabilistic. It doesn’t “think” or “feel” the meaning of words like humans but rather processes language as data, identifying patterns and probabilities to generate plausible responses.
3. ChatGPT Is Only as Good as Your Prompts
As I argued in my Prompting Strategies post, creating well-structured prompts is a fundamental skill, one that can make or break your interaction with conversational AI chatbots such as ChatGPT. The technical term for this practice is “prompt engineering.” Essentially, prompts are the keys to unlocking the capabilities of generative AI; the more refined they are, the better the outcomes the bot produces. Even with the same query, the way we structure our prompt can lead to highly nuanced outcomes.
In my book, I discuss several strategies to help you master this skill. Some of these strategies include writing clear instructions, providing reference text, splitting complex tasks into simpler subtasks, and allowing the model time to “think.” These techniques not only improve the quality of the interactions but also enhance the relevance and accuracy of the responses you receive.
In order to maximize the effectiveness of these strategies, it’s essential to practice and refine your approach to prompt engineering. Over time, this practice will enable you to harness the full potential of AI in a way that complements your specific needs and enhances your projects or educational objectives.
4. ChatGPT Can and Does Hallucinate
One of the critical limitations I discuss in my book is the tendency of AI, like ChatGPT, to deviate from accuracy, “hallucinate,” and produce factual inaccuracies. This is particularly true when it comes to handling factual data and statistical content, where ChatGPT should never be wholly trusted.
This underscores the importance of developing not just your skills in prompt engineering but also your ability to fact-check rigorously. Just because ChatGPT provides an answer does not guarantee its truthfulness. Make it a habit to fact-check all critical information independently, cross-verifying it against trusted sources to ensure its accuracy.
The same principle applies to references and citations. If you request citations or references from ChatGPT, be aware that it can sometimes provide erroneous references and non-existent citations. Therefore, I recommend that you always double-check these citations yourself. Use established academic databases or directly consult primary sources to confirm the validity of the information provided. This extra step is crucial for maintaining the integrity of your work and ensuring that the information you present is reliable and substantiated.
5. Generative AI is Here to Stay
Every time I discuss the staying power of generative AI on my blog, I inevitably receive feedback from some educators who dismiss it as just another passing trend. Interestingly, this type of skepticism often comes from those who have witnessed many such “fads” come and go, and it’s understandable. Change can be unsettling, especially when it challenges the comfortable routines and norms we are accustomed to.
However, embracing change can be exactly what we need to stay relevant and adaptive in our rapidly evolving world. ChatGPT, for instance, has been around for over two years now and has made significant inroads into nearly every aspect of our lives. Ignoring its presence means missing out on a powerful tool that our students are already using in various ways, including some that are less than ideal.
As educators, we have a duty to engage with this new technology. By understanding and integrating generative AI into our teaching, we can harness its educational potential, optimize our instructional practices, and ultimately enhance student learning. Embracing generative AI isn’t just about keeping up with tech trends—it’s about proactively shaping how these tools influence education and ensuring that we’re using them to benefit our students in the best ways possible.
6. Generative AI is a Branch of AI
Many people still use the terms artificial intelligence (AI) and generative artificial intelligence interchangeably, but it’s important to distinguish between the two. As defined in my book, ChatGPT for Teachers, AI is a branch of computer science that develops machines and systems capable of performing complex tasks traditionally requiring human intelligence. AI operates by analyzing data to detect patterns, automate processes, and generate predictions, and has been a field of study since the 1950s.
Generative AI, however, is a specialized subset of AI that focuses on creating synthesized content such as text, images, videos, music, and code in response to user-generated prompts. Conversational AI chatbots like ChatGPT, Gemini, Copilot, and Perplexity are examples of generative AI, showcasing its unique ability to produce new content based on specific inputs.
