The Art of the Pause: Why Effective Writing Requires Unplugged Thinking

I asked Google Gemini, "Write a blog post for WORDS ON THE LINE by Philip Vassallo."  In less than five seconds, it replied, "Here is a draft for a new blog post tailored to the tone and philosophy of your blog." After the line break below, I have left intact the formatting and paragraphing, and I have not changed a word or punctuation mark of this post, including the title.

I have done this thought experiment for three reasons: First, I was curious about AI'sselection of content. Here I would give it a C because I have discussed the points it mentions in so many words repeatedly over the past 22 years of writing WORDS ON THE LINE. Second, I was interested in seeing how close AI comes to imitating my sense of structure (I would give it an A, as I have been teaching this organization strategy for decades) and my approach to style (here it gets a B since I would be more concise). Third, and most important, I wanted to know if AI could add value to WORDS ON THE LINE for my readers. To this point, I am undecided. But I promise not to fade away in the face of AI. To be continued.

The AI Revolution in Education, Part 3: Further Suggestions

No single policy will resolve the AI challenge in education. As technology changes, educators will have to change with it. Here are six evergreen practices that should be at the heart of instruction with AI.

1. Explain  the purpose of every assignment. If students are to practice independent writing, unrestricted AI assistance may defeat the purpose. If they are evaluating arguments, AI might provide useful material to critique.
2. Require students to disclose their AI use. This practice is no different from requiring students to cite their sources. In using AI, what did they request, accept, reject, or revise? Making AI use visible turns it from a hidden shortcut into an accountable part of the learning process.
3. Make AI itself an object of study. Give students an AI-generated response and ask them to find its factual, logical, or stylistic weaknesses. Through such an exercise, students can learn an important lesson: fluent language is not necessarily reliable knowledge.
4. Vary assessments. Discussions, presentations, drafts, demonstrations, reflections, collaborative work, in-class writing, and other teaching modes can provide evidence that a final product alone cannot.
5. Invest in teacher learning. Teachers need training to cultivate their AI knowledge, ethical understanding, pedagogical skills, and continuing professional competence. 
6. Create thinking spaces. The education system should place independent thinking at the heart of all learning. If students are to think for themselves, they must have facilitators ushering them into spaces, physical and otherwise, where AI answers are not immediate. Students and teachers alike must prize the uniquely precious human gift of abstract thinking. 

Not every problem needs an instant answer, nor does every sentence need technological assistance. Sometimes the questions are more important than the answers. Learning can be hard, and that's a good thing. Instead of removing all difficulties, add new challenges. Educators should not keep AI outside the classroom at all costs, nor to put it everywhere simply because it is available. They should include it where strengthens learning and exclude it where it replaces learning.

The AI Revolution in Education, Part 2: The Response

If educators cannot simply eliminate artificial intelligence from the classroom, how should they approach it? So far, we have devised some practical responses.

1. Start with the mindset that AI is an assistant, not the source of knowledge. Using AI purposefully will make students more responsible for learning. When they obtain information based on their results-driven, informed, and focused prompts, they will see AI as another powerful resource, not as the center of learning. 

2. Teach students to question AI rather than simply use it. Students can compare an AI response with reliable sources, identify errors, challenge its reasoning, or ask why one answer is more convincing than another. This approach brings critical thinking into AI literacy.

3. Recreate assignments. Instead of asking students only for a finished essay, teachers can examine outlines, drafts, revisions, sources, and reflections. The process becomes part of the learning.

These approaches suggest that educators do not necessarily have to choose between AI and no AI. They can instead ask whether a particular use of AI advances the learning objective. That question matters because the appropriate use of AI will differ by age, subject, assignment, and purpose, and a science laboratory, a literature discussion, and a college research paper should not necessarily have the same AI rules.

The essential shift is from asking, "How do we stop students from using AI?" to asking, "What do we want students to learn, and what evidence will show that they learned it?" AI may change the answers to those questions, but it must not eliminate those questions.

The AI Revolution in Education, Part 1: The Problem

Artificial intelligence has created dilemmas for educators with no simple solution. Should schools ban it, permit it, or regulate it? Each choice creates at least six problems:

1. Enhancing learning. If AI writes a student's paper, what has the student learned? Evidence suggests that students recall less of the AI-generated content in their assignments, suggesting they do not study the new information.
2. Legitimizing Writing. If AI can produce a competent essay in seconds, why should students struggle to write one? This question implies that writing may go the way of the slide rule.
3. Equalizing access. If students do not have equal access to technology and sophisticated AI tools, will the have-nots have  the knowledge required to use them effectively? AI can therefore create new educational inequalities, even as it promises greater access to information.
4. Protecting privacy. Might students and teachers enter personal information or student work into systems whose data practices they do not fully understand? UNESCO identifies privacy, human agency, accountability, and equity as important considerations in educational AI.
5. Assigning responsibility. If AI provides a wrong answer, who is accountable for accepting responsibility? This point is no small one because content verification can be untraceable.
6. Engendering relevance. What should education preserve when technology changes what people can do? Answering this question is not only at the heart of education but a major factor for giving people worldwide relevant employment.

AI technology is changing every imaginable industry from dairy to diamonds, from medicine to electricity. The challenge is not simply deciding whether AI belongs in education, but which parts of education should remain unmistakably human.